Cloud Native App

Contents
Architecting Cloud-Native .NET Apps for Azure

Introduction to cloud-native applications
Defining Cloud Native
Candidate apps for Cloud Native
Introducing the eShopOnContainers reference app
Mapping eShopOnContainers to Azure Services
Deploying eShopOnContainers to Azure
Centralized configuration
Scaling cloud-native .NET applications
Leveraging containers and orchestrators
Leveraging serverless functions
Combining containers and serverless approaches
Deploying containers in Azure
Scaling containers and serverless applications
Other deployment options
Cloud-native communication patterns
Front-end client communication
Service to service communication
gRPC
Service Mesh communication infrastructure
Distributed data
Relational vs. NoSQL data
Caching in a cloud-native application
Elasticsearch in Azure
Cloud-native resiliency
Application resiliency patterns
Cloud infrastructure resiliency with Azure
Resilient communication
Monitoring and health
Observability patterns
Logging with Elastic Stack
Monitoring in Azure Kubernetes Services
Azure Monitor
Identity
Authentication and authorization in cloud-native apps
Azure Active Directory
Identity Server
Security
Azure Security for cloud-native apps
DevOps
Feature flags
Infrastructure as Code
Cloud Native Application Bundles
Summary
Architecting Cloud Native .NET Applications for
Azure
7/1/2020 • 2 minutes to read • Edit Online
EDITION v.1.0
PUBLISHED BY
Microsoft Developer Division, .NET, and Visual Studio product teams
A division of Microsoft Corporation
One Microsoft Way
Redmond, Washington 98052-6399
Copyright © 2020 by Microsoft Corporation
All rights reserved. No part of the contents of this book may be reproduced or transmitted in any form or by any
means without the written permission of the publisher.
This book is provided "as-is" and expresses the author's views and opinions. The views, opinions, and information
expressed in this book, including URL and other Internet website references, may change without notice.
Some examples depicted herein are provided for illustration only and are fictitious. No real association or
connection is intended or should be inferred.
Microsoft and the trademarks listed at https://www.microsoft.com on the "Trademarks" webpage are trademarks of
the Microsoft group of companies.
Mac and macOS are trademarks of Apple Inc.
The Docker whale logo is a registered trademark of Docker, Inc. Used by permission.
All other marks and logos are property of their respective owners.
Authors:
Rob Vettor , Principal Cloud System Architect/IP Architect - thinkingincloudnative.com, Microsoft

Steve "ardalis" Smith , Software Architect and Trainer - Ardalis.com
Participants and Reviewers:
Cesar De la Torre , Principal Program Manager, .NET team, Microsoft

Nish Anil , Senior Program Manager, .NET team, Microsoft
Jeremy Likness , Senior Program Manager, .NET team, Microsoft
Cecil Phillip , Senior Cloud Advocate, Microsoft
Editors:
Maira Wenzel , Program Manager, .NET team, Microsoft
Version
This guide has been written to cover .NET Core 3.1 version along with many additional updates related to the
same “wave” of technologies (that is, Azure and additional third-party technologies) coinciding in time with the
.NET Core 3.1 release.
Who should use this guide

The audience for this guide is mainly developers, development leads, and architects who are interested in learning
how to build applications designed for the cloud.
A secondary audience is technical decision-makers who plan to choose whether to build their applications using a
cloud-native approach.
How you can use this guide

This guide begins by defining cloud native and introducing a reference application built using cloud-native
principles and technologies. Beyond these first two chapters, the rest of the book is broken up into specific chapters
focused on topics common to most cloud-native applications. You can jump to any of these chapters to learn about
cloud-native approaches to:
Data and data access
Communication patterns
Scaling and scalability
Application resiliency
Identity and security
DevOps
This guide is available both in PDF form and online. Feel free to forward this document or links to its online version
to your team to help ensure common understanding of these topics. Most of these topics benefit from a consistent
understanding of the underlying principles and patterns, as well as the trade-offs involved in decisions related to
these topics. Our goal with this document is to equip teams and their leaders with the information they need to
make well-informed decisions for their applications' architecture, development, and hosting.
Send your feedback

This book and related samples are constantly evolving, so your feedback is welcomed! If you have comments about
how this book can be improved, use the feedback section at the bottom of any page built on GitHub issues.
NE XT
Introduction to cloud-native applications
Another day, at the office, working on "the next big thing."

Your cellphone rings. It's your friendly recruiter - the one who calls you twice a day about new jobs.
But this time it's different: Start-up, equity, and plenty of funding.
The mention of the cloud and cutting-edge technology pushes you over the edge.
Fast forward a few weeks and you're now a new employee in a design session architecting a major eCommerce
application. You're going to compete with the leading eCommerce sites.
How will you build it?
If you follow the guidance from past 15 years, you'll most likely build the system shown in Figure 1.1.
Figure 1-1 . Traditional monolithic design

You construct a large core application containing all of your domain logic. It includes modules such as Identity,
Catalog, Ordering, and more. The core app communicates with a large relational database. The core exposes
functionality via an HTML interface.
Congratulations! You just created a monolithic application.
Not all is bad. Monoliths offer some distinct advantages. For example, they're straightforward to...
build
test
deploy
troubleshoot
scale
Many successful apps that exist today were created as monoliths. The app is a hit and continues to evolve, iteration
after iteration, adding more and more functionality.
At some point, however, you begin to feel uncomfortable. You find yourself losing control of the application. As
time goes on, the feeling becomes more intense and you eventually enter a state known as the Fear Cycle .
The app has become so overwhelmingly complicated that no single person understands it.
You fear making changes - each change has unintended and costly side effects.
New features/fixes become tricky, time-consuming, and expensive to implement.
Each release as small as possible and requires a full deployment of the entire application.
One unstable component can crash the entire system.
New technologies and frameworks aren't an option.
It's difficult to implement agile delivery methodologies.
Architectural erosion sets in as the code base deteriorates with never-ending "special cases."
The consultants tell you to rewrite it.
Many organizations have addressed the monolithic fear cycle by adopting a cloud-native approach to building
systems. Figure 1-2 shows the same system built applying cloud-native techniques and practices.
Figure 1-2 . Cloud-native design

Note how the application is decomposed across a set of small isolated microservices. Each service is self-contained
and encapsulates its own code, data, and dependencies. Each is deployed in a software container and managed by
a container orchestrator. Instead of a large relational database, each service owns it own datastore, the type of
which vary based upon the data needs. Note how some services depend on a relational database, but other on
NoSQL databases. One service stores its state in a distributed cache. Note how all traffic routes through an API
Gateway service that is responsible for directing traffic to the core back-end services and enforcing many cross-
cutting concerns. Most importantly, the application takes full advantage of the scalability, availability, and resiliency
features found in modern cloud platforms.
Cloud-native computing
Hmm... We just used the term, Cloud Native. Your first thought might be, "What exactly does that mean?" Another
industry buzzword concocted by software vendors to market more stuff?"
Fortunately it's far different and hopefully this book will help convince you.
Within a short time, cloud native has become a driving trend in the software industry. It's a new way to think about
building large, complex systems, an approach that takes full advantage of modern software development practices,
technologies, and cloud infrastructure. The approach changes the way you design, implement, deploy, and
operationalize systems.
Unlike the continuous hype that drives our industry, cloud native is for-real. Consider the Cloud Native Computing
Foundation (CNCF), a consortium of over 300 major corporations with a charter to make cloud-native computing
ubiquitous across technology and cloud stacks. As one of the most influential open-source groups, it hosts many of
the fastest-growing open source-projects in GitHub. They include projects such as Kubernetes, Prometheus, Helm,
Envoy, and gRPC.
The CNCF fosters an ecosystem of open-source and vendor-neutrality. Following that lead, this book presents
cloud-native principles, patterns, and best practices that are technology agnostic. At the same time, we discuss the
services and infrastructure available in the Microsoft Azure cloud for constructing cloud-native systems.
So, what exactly is Cloud Native? Sit back, relax, and let us help you explore this new world.
PR EVIO U S NE XT
Defining cloud native
Stop what you're doing and text ten of your colleagues. Ask them to define the term "Cloud Native". Good chance
you'll get ten different answers.
Cloud native is all about changing the way you think about constructing critical business systems.
Cloud-native systems are designed to embrace rapid change, large scale, and resilience.
The Cloud Native Computing Foundation provides an official definition:
Cloud-native technologies empower organizations to build and run scalable applications in modern, dynamic
environments such as public, private, and hybrid clouds. Containers, service meshes, microservices, immutable
infrastructure, and declarative APIs exemplify this approach.
These techniques enable loosely coupled systems that are resilient, manageable, and observable. Combined
with robust automation, they allow engineers to make high-impact changes frequently and predictably with
minimal toil.
Applications have become increasingly complex with users demanding more and more. Users expect rapid
responsiveness, innovative features, and zero downtime. Performance problems, recurring errors, and the inability
to move fast are no longer acceptable. They'll easily move to your competitor.
Cloud native is much about speed and agility. Business systems are evolving from enabling business capabilities to
weapons of strategic transformation, accelerating business velocity and growth. It's imperative to get ideas to
market immediately.
Here are some companies who have implemented these techniques. Think about the speed, agility, and scalability
they've achieved.
C O M PA N Y EXP ERIEN C E
Netflix Has 600+ services in production. Deploys hundred times per

day.
Uber Has 1,000+ services in production. Deploys several thousand

times each week.
WeChat Has 300+ services in production. Deploys 1,000 times a day.
As you can see, Netflix, Uber, and WeChat expose systems that consist of hundreds of independent microservices.
This architectural style enables them to rapidly respond to market conditions. They can instantaneously update
small areas of a live, complex application, and individually scale those areas as needed.
The speed and agility of cloud native come about from a number of factors. Foremost is cloud infrastructure. Five
additional foundational pillars shown in Figure 1-3 also provide the bedrock for cloud-native systems.
Figure 1-3 . Cloud-native foundational pillars
Let's take some time to better understand the significance of each pillar.
The cloud…
Cloud-native systems take full advantage of the cloud service model.
Designed to thrive in a dynamic, virtualized cloud environment, these systems make extensive use of Platform as a
Service (PaaS) compute infrastructure and managed services. They treat the underlying infrastructure as
disposable - provisioned in minutes and resized, scaled, moved, or destroyed on demand – via automation.
Consider the widely accepted DevOps concept of Pets vs. Cattle. In a traditional data center, servers are treated as
Pets: a physical machine, given a meaningful name, and cared for. You scale by adding more resources to the same
machine (scaling up). If the server becomes sick, you nurse it back to health. Should the server become
unavailable, everyone notices.
The Cattle service model is different. You provision each instance as a virtual machine or container. They're identical
and assigned a system identifier such as Service-01, Service-02, and so on. You scale by creating more of them
(scaling out). When one becomes unavailable, nobody notices.
The cattle model embraces immutable infrastructure. Servers aren't repaired or modified. If one fails or requires
updating, it's destroyed and a new one is provisioned – all done via automation.
Cloud-native systems embrace the Cattle service model. They continue to run as the infrastructure scales in or out
with no regard to the machines upon which they're running.
The Azure cloud platform supports this type of highly elastic infrastructure with automatic scaling, self-healing, and
monitoring capabilities.
Modern design
How would you design a cloud-native app? What would your architecture look like? To what principles, patterns,
and best practices would you adhere? What infrastructure and operational concerns would be important?
The Twelve -Factor Application
A widely accepted methodology for constructing cloud-based applications is the Twelve-Factor Application. It
describes a set of principles and practices that developers follow to construct applications optimized for modern
cloud environments. Special attention is given to portability across environments and declarative automation.
While applicable to any web-based application, many practitioners consider Twelve-Factor as a solid foundation for
building cloud-native apps. Systems built upon these principles can deploy and scale rapidly and add features to
react quickly to market changes.
The following table highlights the Twelve-Factor methodology:
FA C TO R EXP L A N AT IO N
1 Code Base A single code base for each

microservice, stored in its own
repository. Tracked with version control,
it can deploy to multiple environments
(QA, Staging, Production).
2 Dependencies Each microservice isolates and packages

its own dependencies, embracing
changes without impacting the entire
system.
3 Configurations Configuration information is moved out

of the microservice and externalized
through a configuration management
tool outside of the code. The same
deployment can propagate across
environments with the correct
configuration applied.
4 Backing Services Ancillary resources (data stores, caches,

message brokers) should be exposed
via an addressable URL. Doing so
decouples the resource from the
application, enabling it to be
interchangeable.
5 Build, Release, Run Each release must enforce a strict

separation across the build, release, and
run stages. Each should be tagged with
a unique ID and support the ability to
roll back. Modern CI/CD systems help
fulfill this principle.
6 Processes Each microservice should execute in its

own process, isolated from other
running services. Externalize required
state to a backing service such as a
distributed cache or data store.
7 Port Binding Each microservice should be self-

contained with its interfaces and
functionality exposed on its own port.
Doing so provides isolation from other
microservices.
8 Concurrency Services scale out across a large

number of small identical processes
(copies) as opposed to scaling-up a
single large instance on the most
powerful machine available.
9 Disposability Service instances should be disposable,

favoring fast startups to increase
scalability opportunities and graceful
shutdowns to leave the system in a
correct state. Docker containers along
with an orchestrator inherently satisfy
this requirement.
FA C TO R EXP L A N AT IO N
10 Dev/Prod Parity Keep environments across the

application lifecycle as similar as
possible, avoiding costly shortcuts.
Here, the adoption of containers can
greatly contribute by promoting the
same execution environment.
11 Logging Treat logs generated by microservices

as event streams. Process them with an
event aggregator and propagate the
data to data-mining/log management
tools like Azure Monitor or Splunk and
eventually long-term archival.
12 Admin Processes Run administrative/management tasks

as one-off processes. Tasks can include
data cleanup and pulling analytics for a
report. Tools executing these tasks
should be invoked from the production
environment, but separately from the
application.
In the book, Beyond the Twelve-Factor App, author Kevin Hoffman details each of the original 12 factors (written in
2011). Additionally, he discusses three additional factors that reflect today's modern cloud application design.
N EW FA C TO R EXP L A N AT IO N
13 API First Make everything a service. Assume

your code will be consumed by a front-
end client, gateway, or another service.
14 Telemetry On a workstation, you have deep

visibility into your application and its
behavior. In the cloud, you don't. Make
sure your design includes the collection
of monitoring, domain-specific, and
health/system data.
15 Authentication/ Authorization Implement identity from the start.

Consider RBAC (role-based access
control) features available in public
clouds.
We'll refer to many of the 12+ factors in this chapter and throughout the book.
Critical Design Considerations
Beyond the guidance provided from the twelve-factor methodology, there are several critical design decisions you
must make when constructing distributed systems.
Communication
How will front-end client applications communicate with backed-end core services? Will you allow direct
communication? Or, might you abstract the back-end services with a gateway façade that provides flexibility,
control, and security?
How will back-end core services communicate with each other? Will you allow direct HTTP calls that lead to
coupling and impact performance and agility? Or might you consider decoupled messaging with queue and topic
technologies?
Communication is covered in detail Chapter 4, Cloud-Native Communication Patterns.
Resiliency
A microservices architecture moves your system from in-process to out-of-process network communication. In a
distributed architecture, what happens when Service B isn't responding to a network call from Service A? Or, what
happens when Service C becomes temporarily unavailable and other services calling it are blocked?
Resiliency is covered in detail Chapter 6, Cloud-Native Resiliency.
Distributed Data
By design, each microservice encapsulates its own data, exposing operations via its public interface. If so, how do
you query data or implement a transaction across multiple services?
Distributed data is covered in detail Chapter 5, Cloud-Native Data Patterns.
Identity
How will your service identify who is accessing it and what permissions they have?
Identity is covered in detail Chapter 8, Identity.
Microservices
Cloud-native systems embrace microservices, a popular architectural style for constructing modern applications.
Built as a distributed set of small, independent services that interact through a shared fabric, microservices share
the following characteristics:
Each implements a specific business capability within a larger domain context.
Each is developed autonomously and can be deployed independently.
Each is self-contained encapsulating its own data storage technology (SQL, NoSQL) and programming
platform.
Each runs in its own process and communicates with others using standard communication protocols such
as HTTP/HTTPS, WebSockets, or AMQP.
They compose together to form an application.
Figure 1-4 contrasts a monolithic application approach with a microservices approach. Note how the monolith is
composed of a layered architecture, which executes in a single process. It typically consumes a relational database.
The microservice approach, however, segregates functionality into independent services that include logic and
data. Each microservice hosts its own datastore.
Figure 1-4. Monolithic deployment versus microservices
Note how microservices promote the "One Codebase, One Application" principle from the Twelve-Factor
Application, discussed earlier in the chapter.
Factor #1 specifies "A single codebase for each microservice, stored in its own repository. Tracked with version
control, it can deploy to multiple environments."
Why microservices?
Microservices provide agility.
Earlier in the chapter, we compared an eCommerce application built as a monolith to that with microservices. In the
example, we saw some clear benefits:
Each microservice has an autonomous lifecycle and can evolve independently and deploy frequently. You
don't have to wait for a quarterly release to deploy a new features or update. You can update a small area of
a complex application with less risk of disrupting the entire system.
Each microservice can scale independently. Instead of scaling the entire application as a single unit, you
scale out only those services that require more processing power or network bandwidth. This fine-grained
approach to scaling provides for greater control of your system and helps to reduce overall costs as you
scale portions of your system, not everything.
An excellent reference guide for understanding microservices is .NET Microservices: Architecture for Containerized
.NET Applications. The book deep dives into microservices design and architecture. It's a companion for a full-stack
microservice reference architecture available as a free download from Microsoft.
Developing microservices
Microservices can be created with any modern development platform.
The Microsoft .NET Core platform is an excellent choice. Free and open source, it has many built-in features to
simplify microservice development. .NET Core is cross-platform. Applications can be built and run on Windows,
macOS, and most flavors of Linux.
.NET Core is highly performant and has scored well in comparison to Node.js and other competing platforms.
Interestingly, TechEmpower conducted an extensive set of performance benchmarks across many web application
platforms and frameworks. .NET Core scored in the top 10 - well above Node.js and other competing platforms.
.NET Core is maintained by Microsoft and the .NET community on GitHub.
Containers
Nowadays, it's natural to hear the term container mentioned in any conversation concerning cloud native. In the
book, Cloud Native Patterns, author Cornelia Davis observes that, "Containers are a great enabler of cloud-native
software." The Cloud Native Computing Foundation places microservice containerization as the first step in their
Cloud-Native Trail Map - guidance for enterprises beginning their cloud-native journey.
Containerizing a microservice is simple and straightforward. The code, its dependencies, and runtime are packaged
into a binary called a container image. Images are stored in a container registry, which acts as a repository or
library for images. A registry can be located on your development computer, in your data center, or in a public
cloud. Docker itself maintains a public registry via Docker Hub. The Azure cloud features a container registry to
store container images close to the cloud applications that will run them.
When needed, you transform the image into a running container instance. The instance runs on any computer that
has a container runtime engine installed. You can have as many instances of the containerized service as needed.
Figure 1-5 shows three different microservices, each in its own container, running on a single host.
Figure 1-5 . Multiple containers running on a container host

Note how each container maintains its own set of dependencies and runtime, which can be different. Here, we see
different versions of the Product microservice running on the same host. Each container shares a slice of the
underlying host operating system, memory, and processor, but is isolated from one another.
Note how well the container model embraces the "Dependencies" principle from the Twelve-Factor Application.
Factor #2 specifies that "Each microservice isolates and packages its own dependencies, embracing changes
without impacting the entire system."
Containers support both Linux and Windows workloads. The Azure cloud openly embraces both. Interestingly, it's
Linux, not Windows Server, that has become the most popular operating system in Azure.
While several container vendors exist, Docker has captured the lion's share of the market. The company has been
driving the software container movement. It has become the de facto standard for packaging, deploying, and
running cloud-native applications.
Why containers?
Containers provide portability and guarantee consistency across environments. By encapsulating everything into a
single package, you isolate the microservice and its dependencies from the underlying infrastructure.
You can deploy that same container in any environment that has the Docker runtime engine. Containerized
workloads also eliminate the expense of pre-configuring each environment with frameworks, software libraries,
and runtime engines.
By sharing the underlying operating system and host resources, containers have a much smaller footprint than a
full virtual machine. The smaller size increases the density, or number of microservices, that a given host can run at
one time.
Container orchestration
While tools such as Docker create images and run containers, you also need tools to manage them. Container
management is done with a special software program called a container orchestrator. When operating at scale,
container orchestration is essential.
Figure 1-6 shows management tasks that container orchestrators provide.
Figure 1-6 . What container orchestrators do

The following table describes common orchestration tasks.
TA SK S EXP L A N AT IO N
Scheduling Automatically provision container instances.
Affinity/anti-affinity Provision containers nearby or far apart from each other,

helping availability and performance.
Health monitoring Automatically detect and correct failures.
Failover Automatically reprovision failed instance to healthy machines.
Scaling Automatically add or remove container instance to meet

demand.
Networking Manage a networking overlay for container communication.
Service Discovery Enable containers to locate each other.
Rolling Upgrades Coordinate incremental upgrades with zero downtime

deployment. Automatically roll back problematic changes.
Note how orchestrators embrace the disposability and concurrency principles from the Twelve-Factor Application,
discussed earlier in the chapter.
Factor #9 specifies that "Service instances should be disposable, favoring fast startups to increase scalability
opportunities and graceful shutdowns to leave the system in a correct state. Docker containers along with an
orchestrator inherently satisfy this requirement."
Factor #8 specifies that "Services scale out across a large number of small identical processes (copies) as
opposed to scaling-up a single large instance on the most powerful machine available."
While several container orchestrators exist, Kubernetes has become the de facto standard for the cloud-native
world. It's a portable, extensible, open-source platform for managing containerized workloads.
You could host your own instance of Kubernetes, but then you'd be responsible for provisioning and managing its
resources - which can be complex. The Azure cloud features Kubernetes as a managed service, Azure Kubernetes
Service (AKS). A managed service allows you to fully leverage its features, without having to install and maintain it.
Azure Kubernetes Services is covered in detail Chapter 2, Scaling Cloud-Native Applications.
Backing services
Cloud-native systems depend upon many different ancillary resources, such as data stores, message brokers,
monitoring, and identity services. These services are known as backing services.
Figure 1-7 shows many common backing services that cloud-native systems consume.
Figure 1-7 . Common backing services

Backing services promote the "Statelessness" principle from the Twelve-Factor Application, discussed earlier in the
chapter.
Factor #6 specifies that, "Each microservice should execute in its own process, isolated from other running
services. Externalize required state to a backing service such as a distributed cache or data store."
You could host your own backing services, but then you'd be responsible for licensing, provisioning, and managing
those resources.
Cloud providers offer a rich assortment of managed backing services. Instead of owning the service, you simply
consume it. The provider operates the resource at scale and bears the responsibility for performance, security, and
maintenance. Monitoring, redundancy, and availability are built into the service. Providers fully support their
managed services - open a ticket and they fix your issue.
Cloud-native systems favor managed backing services from cloud vendors. The savings in time and labor are
great. The operational risk of hosting your own and experiencing trouble can get expensive fast.
A best practice is to treat a backing service as an attached resource, dynamically bound to a microservice with
information (a URL and credentials) stored in an external configuration. This guidance is spelled out in the Twelve-
Factor Application, discussed earlier in the chapter.
Factor #4 specifies that backing services "should be exposed via an addressable URL. Doing so decouples the
resource from the application, enabling it to be interchangeable."
Factor #3 specifies that "Configuration information is moved out of the microservice and externalized through
a configuration management tool outside of the code."
With this pattern, a backing service can be attached and detached without code changes. You might promote a
microservice from QA to a staging environment. You update the microservice configuration to point to the backing
services in staging and inject the settings into your container through an environment variable.
Cloud vendors provide APIs for you to communicate with their proprietary backing services. These libraries
encapsulate the plumbing and complexity. Communicating directly with these APIs will tightly couple your code to
the backing service. It's a better practice to insulate the implementation details of the vendor API. Introduce an
intermediation layer, or intermediate API, exposing generic operations to your service code. This loose coupling
enables you to swap out one backing service for another or move your code to a different public cloud without
having to make changes to the mainline service code.
Backing services are discussed in detail Chapter 5, Cloud-Native Data Patterns, and Chapter 4, Cloud-Native
Communication Patterns.
Automation
As you've seen, cloud-native systems embrace microservices, containers, and modern system design to achieve
speed and agility. But, that's only part of the story. How do you provision the cloud environments upon which these
systems run? How do you rapidly deploy app features and updates? How do you round out the full picture?
Enter the widely accepted practice of Infrastructure as Code, or IaC.
With IaC, you automate platform provisioning and application deployment. You essentially apply software
engineering practices such as testing and versioning to your DevOps practices. Your infrastructure and
deployments are automated, consistent, and repeatable.
Automating infrastructure
Tools like Azure Resource Manager, Terraform, and the Azure CLI, enable you to declaratively script the cloud
infrastructure you require. Resource names, locations, capacities, and secrets are parameterized and dynamic. The
script is versioned and checked into source control as an artifact of your project. You invoke the script to provision
a consistent and repeatable infrastructure across system environments, such as QA, staging, and production.
Under the hood, IaC is idempotent, meaning that you can run the same script over and over without side effects. If
the team needs to make a change, they edit and rerun the script. Only the updated resources are affected.
In the article, What is Infrastructure as Code, Author Sam Guckenheimer describes how, "Teams who implement IaC
can deliver stable environments rapidly and at scale. Teams avoid manual configuration of environments and
enforce consistency by representing the desired state of their environments via code. Infrastructure deployments
with IaC are repeatable and prevent runtime issues caused by configuration drift or missing dependencies. DevOps
teams can work together with a unified set of practices and tools to deliver applications and their supporting
infrastructure rapidly, reliably, and at scale."
Automating deployments
The Twelve-Factor Application, discussed earlier, calls for separate steps when transforming completed code into a
running application.
Factor #5 specifies that "Each release must enforce a strict separation across the build, release and run stages.
Each should be tagged with a unique ID and support the ability to roll back."
Modern CI/CD systems help fulfill this principle. They provide separate deployment steps and help ensure
consistent and quality code that's readily available to users.
Figure 1-8 shows the separation across the deployment process.
Figure 1-8 . Deployment steps in a CI/CD Pipeline
In the previous figure, pay special attention to separation of tasks.
The developer constructs a feature in their development environment, iterating through what is called the "inner
loop" of code, run, and debug. When complete, that code is pushed into a code repository, such as GitHub, Azure
DevOps, or BitBucket.
The push triggers a build stage that transforms the code into a binary artifact. The work is implemented with a
Continuous Integration (CI) pipeline. It automatically builds, tests, and packages the application.
The release stage picks up the binary artifact, applies external application and environment configuration
information, and produces an immutable release. The release is deployed to a specified environment. The work is
implemented with a Continuous Delivery(CD) pipeline. Each release should be identifiable. You can say, "This
deployment is running Release 2.1.1 of the application."
Finally, the released feature is run in the target execution environment. Releases are immutable meaning that any
change must create a new release.
Applying these practices, organizations have radically evolved how they ship software. Many have moved from
quarterly releases to on-demand updates. The goal is to catch problems early in the development cycle when
they're less expensive to fix. The longer the duration between integrations, the more expensive problems become
to resolve. With consistency in the integration process, teams can commit code changes more frequently, leading
to better collaboration and software quality.
Azure Pipelines
The Azure cloud includes a new CI/CD service entitled Azure Pipelines, which is part of the Azure DevOps offering
shown in Figure 1-9.
Figure 1-9 . Azure DevOps offerings

Azure Pipelines is a cloud service that combines continuous integration (CI) and continuous delivery (CD). You can
automatically test, build, and ship your code to any target.
You define your pipeline in code in a YAML file alongside the rest of the code for your app.
The pipeline is versioned with your code and follows the same branching structure.
You get validation of your changes through code reviews in pull requests and branch build policies.
Every branch you use can customize the build policy by modifying the azure-pipelines.yml file.
The pipeline file is checked into version control and can be investigated if there's a problem.
The Azure Pipelines service supports most Git providers and can generate deployment pipelines for applications
written on the Linux, macOS, or Windows platforms. It includes support for Java, .NET, JavaScript, Python, PHP, Go,
XCode, and C++.
PR EVIO U S NE XT
Candidate apps for cloud native
Look at the apps in your portfolio. How many of them qualify for a cloud-native architecture? All of them? Perhaps
some?
Applying a cost/benefit analysis, there's a good chance that most wouldn't support the hefty price tag required to
be cloud native. The cost of being cloud native would far exceed the business value of the application.
What type of application might be a candidate for cloud native?
Strategic enterprise system that need to constantly evolve business capabilities/features
An application that requires a high release velocity - with high confidence
A system with where individual features must release without a full redeployment of the entire system
An application developed by teams with expertise in different technology stacks
An application with components that must scale independently
Then there are legacy systems. While we'd all like to build new applications, we're often responsible for
modernizing legacy workloads that are critical to the business. Over time, a legacy application could be
decomposed into microservices, containerized, and ultimately "replatformed" into a cloud-native architecture.
Modernizing legacy apps
The free Microsoft e-book Modernize existing .NET applications with Azure cloud and Windows Containers
provides guidance for migrating on-premises workloads into cloud. Figure 1-10 shows that there isn't a single,
one-size-fits-all strategy for modernizing legacy applications.
Figure 1-10 . Strategies for migrating legacy workloads

Monolithic apps that are non-critical largely benefit from a quick lift-and-shift (Cloud Infrastructure-Ready)
migration. Here, the on-premises workload is rehosted to a cloud-based VM, without changes. This approach uses
the IaaS (Infrastructure as a Service) model. Azure includes several tools such as Azure Migrate, Azure Site
Recovery, and Azure Database Migration Service to make such a move easier. While this strategy can yield some
cost savings, such applications typically weren't architected to unlock and leverage the benefits of cloud computing.
Monolithic apps that are critical to the business oftentimes benefit from an enhanced lift-and-shift (Cloud
Optimized) migration. This approach includes deployment optimizations that enable key cloud services - without
changing the core architecture of the application. For example, you might containerize the application and deploy it
to a container orchestrator, like Azure Kubernetes Services, discussed later in this book. Once in the cloud, the
application could consume other cloud services such as databases, message queues, monitoring, and distributed
caching.
Finally, monolithic apps that perform strategic enterprise functions might best benefit from a Cloud-Native
approach, the subject of this book. This approach provides agility and velocity. But, it comes at a cost of
replatforming, rearchitecting, and rewriting code.
If you and your team believe a cloud-native approach is appropriate, it behooves you to rationalize the decision
with your organization. What exactly is the business problem that a cloud-native approach will solve? How would it
align with business needs?
Rapid releases of features with increased confidence?
Fine-grained scalability - more efficient usage of resources?
Improved system resiliency?
Improved system performance?
More visibility into operations?
Blend development platforms and data stores to arrive at the best tool for the job?
Future-proof application investment?
The right migration strategy depends on organizational priorities and the systems you're targeting. For many, it
may be more cost effective to cloud-optimize a monolithic application or add coarse-grained services to an N-Tier
app. In these cases, you can still make full use of cloud PaaS capabilities like the ones offered by Azure App Service.
Summary
In this chapter, we introduced cloud-native computing. We provided a definition along with the key capabilities that
drive a cloud-native application. We looked at the types of applications that might justify this investment and
effort.
With the introduction behind, we now dive into a much more detailed look at cloud native.
References
Cloud Native Computing Foundation
.NET Microservices: Architecture for Containerized .NET applications
Modernize existing .NET applications with Azure cloud and Windows Containers
Cloud Native Patterns by Cornelia Davis
Beyond the Twelve-Factor Application
What is Infrastructure as Code
Uber Engineering's Micro Deploy: Deploying Daily with Confidence
How Netflix Deploys Code
Overload Control for Scaling WeChat Microservices
PR EVIO U S NE XT
Introducing eShopOnContainers reference app
Microsoft, in partnership with leading community experts, has produced a full-featured cloud-native microservices
reference application, eShopOnContainers. This application is built to showcase using .NET Core and Docker, and
optionally Azure, Kubernetes, and Visual Studio, to build an online storefront.
Figure 2-1 . eShopOnContainers Sample App Screenshot.

Before starting this chapter, we recommend that you download the eShopOnContainers reference application. If
you do so, it should be easier for you to follow along with the information presented.
Features and requirements

Let's start with a review of the application's features and requirements. The eShopOnContainers application
represents an online store that sells various physical products like t-shirts and coffee mugs. If you've bought
anything online before, the experience of using the store should be relatively familiar. Here are some of the basic
features the store implements:
List catalog items
Filter items by type
Filter items by brand
Add items to the shopping basket
Edit or remove items from the basket
Checkout
Register an account
Sign in
Sign out
Review orders
The application also has the following non-functional requirements:
It needs to be highly available and it must scale automatically to meet increased traffic (and scale back down
once traffic subsides).
It should provide easy-to-use monitoring of its health and diagnostic logs to help troubleshoot any issues it
encounters.
It should support an agile development process, including support for continuous integration and deployment
(CI/CD).
In addition to the two web front ends (traditional and Single Page Application), the application must also
support mobile client apps running different kinds of operating systems.
It should support cross-platform hosting and cross-platform development.
Figure 2-2 . eShopOnContainers reference application development architecture.

The eShopOnContainers application is accessible from web or mobile clients that access the application over
HTTPS targeting either the ASP.NET Core MVC server application or an appropriate API Gateway. API Gateways
offer several advantages, such as decoupling back-end services from individual front-end clients and providing
better security. The application also makes use of a related pattern known as Backends-for-Frontends (BFF), which
recommends creating separate API gateways for each front-end client. The reference architecture demonstrates
breaking up the API gateways based on whether the request is coming from a web or mobile client.
The application's functionality is broken up into a number of distinct microservices. There are services responsible
for authentication and identity, listing items from the product catalog, managing users' shopping baskets, and
placing orders. Each of these separate services has its own persistent storage. There's no single master data store
with which all services interact. Instead, coordination and communication between the services is done on an as-
needed basis and through the use of a message bus.
Each of the different microservices is designed differently, based on their individual requirements. This means their
technology stack may differ, although they're all built using .NET Core and designed for the cloud. Simpler services
provide basic Create-Read-Update-Delete (CRUD) access to the underlying data stores, while more advanced
services use Domain-Driven Design approaches and patterns to manage business complexity.
Figure 2-3 . Different kinds of microservices.
Overview of the code

Because it leverages microservices, the eShopOnContainers app includes quite a few separate projects and
solutions in its GitHub repository. In addition to separate solutions and executable files, the various services are
designed to run inside their own containers, both during local development and at runtime in production. Figure 2-
4 shows the full Visual Studio solution, in which the various different projects are organized.
Figure 2-4 . Projects in Visual Studio solution.
The code is organized to support the different microservices, and within each microservice, the code is broken up
into domain logic, infrastructure concerns, and user interface or service endpoint. In many cases, each service's
dependencies can be fulfilled by Azure services in production, as well as alternative options for local development.
Let's examine how the application's requirements map to Azure services.
Understanding microservices
This book focuses on cloud-native applications built using Azure technology. To learn more about microservices
best practices and how to architect microservice-based applications, read the companion book, .NET Microservices:
Architecture for Containerized .NET Applications.
PR EVIO U S NE XT
Mapping eShopOnContainers to Azure Services
Although not required, Azure is well-suited to supporting the eShopOnContainers because the project was built to
be a cloud-native application. The application is built with .NET Core, so it can run on Linux or Windows containers
depending on the Docker host. The application is made up of multiple autonomous microservices, each with its
own data. The different microservices showcase different approaches, ranging from simple CRUD operations to
more complex DDD and CQRS patterns. Microservices communicate with clients over HTTP and with one another
via message-based communication. The application supports multiple platforms for clients as well, since it adopts
HTTP as a standard communication protocol and includes ASP.NET Core and Xamarin mobile apps that run on
Android, iOS, and Windows platforms.
The application's architecture is shown in Figure 2-5. On the left are the client apps, broken up into mobile,
traditional Web, and Web Single Page Application (SPA) flavors. On the right are the server-side components that
make up the system, each of which can be hosted in Docker containers and Kubernetes clusters. The traditional
web app is powered by the ASP.NET Core MVC application shown in yellow. This app and the mobile and web SPA
applications communicate with the individual microservices through one or more API gateways. The API gateways
follow the "backends for front ends" (BFF) pattern, meaning that each gateway is designed to support a given
front-end client. The individual microservices are listed to the right of the API gateways and include both business
logic and some kind of persistence store. The different services make use of SQL Server databases, Redis cache
instances, and MongoDB/CosmosDB stores. On the far right is the system's Event Bus, which is used for
communication between the microservices.
Figure 2-5 . The eShopOnContainers Architecture.

The server-side components of this architecture all map easily to Azure services.
Container orchestration and clustering

The application's container-hosted services, from ASP.NET Core MVC apps to individual Catalog and Ordering
microservices, can be hosted and managed in Azure Kubernetes Service (AKS). The application can run locally on
Docker and Kubernetes, and the same containers can then be deployed to staging and production environments
hosted in AKS. This process can be automated as we'll see in the next section.
AKS provides management services for individual clusters of containers. The application will deploy separate AKS
clusters for each microservice shown in the architecture diagram above. This approach allows each individual
service to scale independently according to its resource demands. Each microservice can also be deployed
independently, and ideally such deployments should incur zero system downtime.
API Gateway
The eShopOnContainers application has multiple front-end clients and multiple different back-end services. There's
no one-to-one correspondence between the client applications and the microservices that support them. In such a
scenario, there may be a great deal of complexity when writing client software to interface with the various back-
end services in a secure manner. Each client would need to address this complexity on its own, resulting in
duplication and many places in which to make updates as services change or new policies are implemented.
Azure API Management (APIM) helps organizations publish APIs in a consistent, manageable fashion. APIM
consists of three components: the API Gateway, and administration portal (the Azure portal), and a developer
portal.
The API Gateway accepts API calls and routes them to the appropriate back-end API. It can also provide additional
services like verification of API keys or JWT tokens and API transformation on the fly without code modifications
(for instance, to accommodate clients expecting an older interface).
The Azure portal is where you define the API schema and package different APIs into products. You also configure
user access, view reports, and configure policies for quotas or transformations.
The developer portal serves as the main resource for developers. It provides developers with API documentation,
an interactive test console, and reports on their own usage. Developers also use the portal to create and manage
their own accounts, including subscription and API key support.
Using APIM, applications can expose several different groups of services, each providing a back end for a particular
front-end client. APIM is recommended for complex scenarios. For simpler needs, the lightweight API Gateway
Ocelot can be used. The eShopOnContainers app uses Ocelot because of its simplicity and because it can be
deployed into the same application environment as the application itself. Learn more about eShopOnContainers,
APIM, and Ocelot.
Another option if your application is using AKS is to deploy the Azure Gateway Ingress Controller as a pod within
your AKS cluster. This allows your cluster to integrate with an Azure Application Gateway, allowing the gateway to
load-balance traffic to the AKS pods. Learn more about the Azure Gateway Ingress Controller for AKS.
Data
The various back-end services used by eShopOnContainers have different storage requirements. Several
microservices use SQL Server databases. The Basket microservice leverages a Redis cache for its persistence. The
Locations microservice expects a MongoDB API for its data. Azure supports each of these data formats.
For SQL Server database support, Azure has products for everything from single databases up to highly scalable
SQL Database elastic pools. Individual microservices can be configured to communicate with their own individual
SQL Server databases quickly and easily. These databases can be scaled as needed to support each separate
microservice according to its needs.
The eShopOnContainers application stores the user's current shopping basket between requests. This is managed
by the Basket microservice that stores the data in a Redis cache. In development, this cache can be deployed in a
container, while in production it can utilize Azure Cache for Redis. Azure Cache for Redis is a fully managed service
offering high performance and reliability without the need to deploy and manage Redis instances or containers on
your own.
The Locations microservice uses a MongoDB NoSQL database for its persistence. During development, the
database can be deployed in its own container, while in production the service can leverage Azure Cosmos DB's API
for MongoDB. One of the benefits of Azure Cosmos DB is its ability to leverage multiple different communication
protocols, including a SQL API and common NoSQL APIs including MongoDB, Cassandra, Gremlin, and Azure Table
Storage. Azure Cosmos DB offers a fully managed and globally distributed database as a service that can scale to
meet the needs of the services that use it.
Distributed data in cloud-native applications is covered in more detail in chapter 5.
Event Bus
The application uses events to communicate changes between different services. This functionality can be
implemented with a variety of implementations, and locally the eShopOnContainers application uses RabbitMQ.
When hosted in Azure, the application would leverage Azure Service Bus for its messaging. Azure Service Bus is a
fully managed integration message broker that allows applications and services to communicate with one another
in a decoupled, reliable, asynchronous manner. Azure Service Bus supports individual queues as well as separate
topics to support publisher-subscriber scenarios. The eShopOnContainers application would leverage topics with
Azure Service Bus to support distributing messages from one microservice to any other microservice that needed
to react to a given message.
Resiliency
Once deployed to production, the eShopOnContainers application would be able to take advantage of several
Azure services available to improve its resiliency. The application publishes health checks, which can be integrated
with Application Insights to provide reporting and alerts based on the app's availability. Azure resources also
provide diagnostic logs that can be used to identify and correct bugs and performance issues. Resource logs
provide detailed information on when and how different Azure resources are used by the application. You'll learn
more about cloud-native resiliency features in chapter 6.
PR EVIO U S NE XT
Deploying eShopOnContainers to Azure
The eShopOnContainers application can be deployed to a variety of Azure platforms. The recommended approach
is to deploy the application to Azure Kubernetes Services (AKS). Helm, a Kubernetes deployment tool, is available to
reduce deployment complexity. Optionally, developers may implement Azure Dev Spaces for Kubernetes to
streamline their development process.
Azure Kubernetes Service

To host eShop in AKS, the first step is to create an AKS cluster. To do so, you might use the Azure portal, which will
walk you through the required steps. You could also create a cluster from the Azure CLI, taking care to enable Role-
Based Access Control (RBAC) and application routing. The eShopOnContainers' documentation details the steps for
creating your own AKS cluster. Once created, you can access and manage the cluster from the Kubernetes
dashboard.
You can now deploy the eShop application to the cluster leveraging Helm and Tiller.
Deploying to Azure Kubernetes Service using Helm

Helm is an application package manager tool that works directly with Kubernetes. It helps you define, install, and
upgrade Kubernetes applications. While simple apps can be deployed to AKS with custom CLI scripts or simple
deployment files, complex apps can contain many Kubernetes objects and benefit from Helm.
Using Helm, applications include text-based configuration files, called Helm charts, which declaratively describe the
application and configuration in Helm packages. Charts use standard YAML-formatted files to describe a related set
of Kubernetes resources. They're versioned alongside the application code they describe. Helm Charts range from
simple to complex depending on the requirements of the installation they describe.
Helm is composed of a command-line client tool, which consumes helm charts and launches commands to a
server component named, Tiller. Tiller communicates with the Kubernetes API to ensure the correct provisioning of
your containerized workloads. Helm is maintained by the Cloud-native Computing Foundation.
The following yaml file presents a Helm template:
apiVersion: v1
kind: Service
metadata:
name: {{ .Values.app.svc.marketing }}
labels:
app: {{ template "marketing-api.name" . }}
chart: {{ template "marketing-api.chart" . }}
release: {{ .Release.Name }}
heritage: {{ .Release.Service }}
spec:
type: {{ .Values.service.type }}
ports:
- port: {{ .Values.service.port }}
targetPort: http
protocol: TCP
name: http
selector:
app: {{ template "marketing-api.name" . }}
release: {{ .Release.Name }}
Note how the template describes a dynamic set of key/value pairs. When the template is invoked, values that
enclosed in curly braces are pulled in from other yaml-based configuration files.
You'll find the eShopOnContainers helm charts in the /k8s/helm folder. Figure 2-6 shows how the different
components of the application are organized into a folder structure used by helm to define and managed
deployments.
Figure 2-6 . The eShopOnContainers helm folder.

Each individual component is installed using a helm install command. eShop includes a "deploy all" script that
loops through and installs the components using their respective helm charts. The result is a repeatable process,
versioned with the application in source control, that anyone on the team can deploy to an AKS cluster with a one-
line script command.
Note that version 3 of Helm officially removes the need for the Tiller server component. More information on
this enhancement can be found here.
Azure Dev Spaces

Cloud-native applications can quickly grow large and complex, requiring significant compute resources to run. In
these scenarios, the entire application can't be hosted on a development machine (especially a laptop). Azure Dev
Spaces is designed to address this problem using AKS. It enables developers to work with a local version of their
services while hosting the rest of the application in an AKS development cluster.
Developers share a running (development) instance in an AKS cluster that contains the entire containerized
application. But they use personal spaces set up on their machine to locally develop their services. When ready,
they test from end-to-end in the AKS cluster - without replicating dependencies. Azure Dev Spaces merges code
from the local machine with services in AKS. Team members can see how their changes will behave in a real AKS
environment. Developers can rapidly iterate and debug code directly in Kubernetes using Visual Studio 2017 or
Visual Studio Code.
In Figure 2-7, you can see that Developer Susie has deployed an updated version of the Bikes microservice into her
dev space. She's then able to test her changes using a custom URL starting with the name of her space
(susie.s.dev.myapp.eus.azds.io).
Figure 2-7 . Developer Susie deploys her own version of the Bikes microservice and tests it.
At the same time, developer John is customizing the Reservations microservice and needs to test his changes. He
deploys his changes to his own dev space without conflicting with Susie's changes as shown in Figure 2-8. John
then tests his changes using his own URL that is prefixed with the name of his space ( john.s.dev.myapp.eus.azds.io).
Figure 2-8 . Developer John deploys his own version of the Reservations microservice and tests it without
conflicting with other developers.
Using Azure Dev Spaces, teams can work directly with AKS while independently changing, deploying, and testing
their changes. This approach reduces the need for separate dedicated hosted environments since every developer
effectively has their own AKS environment. Developers can work with Azure Dev Spaces using its CLI or launch
their application to Azure Dev Spaces directly from Visual Studio. Learn more about how Azure Dev Spaces works
and is configured.
Azure Functions and Logic Apps (Serverless)

The eShopOnContainers sample includes support for tracking online marketing campaigns. An Azure Function is
used to track marketing campaign details for a given campaign ID. Rather than creating a full microservice, a single
Azure Function is simpler and sufficient. Azure Functions have a simple build and deployment model, especially
when configured to run in Kubernetes. Deploying the function is scripted using Azure Resource Manager (ARM)
templates and the Azure CLI. This campaign service isn't customer-facing and invokes a single operation, making it
a great candidate for Azure Functions. The function requires minimal configuration, including a database
connection string data and image base URI settings. You configure Azure Functions in the Azure portal.
PR EVIO U S NE XT
Centralized configuration
Unlike a monolithic app in which everything runs within a single instance, a cloud-native application consists of
independent services distributed across virtual machines, containers, and geographic regions. Managing
configuration settings for dozens of interdependent services can be challenging. Duplicate copies of configuration
settings across different locations is error prone and difficult to manage. Centralized configuration is a critical
requirement for distributed cloud-native applications.
As discussed in Chapter 1, the Twelve-Factor App recommendations require strict separation between code and
configuration. Configuration must be stored externally from the application and read-in as needed. Storing
configuration values as constants or literal values in code is a violation. The same configuration values are often be
used by many services in the same application. Additionally, we must support the same values across multiple
environments, such as dev, testing, and production. The best practice is store them in a centralized configuration
store.
The Azure cloud presents several great options.
Azure App Configuration

Azure App Configuration is a fully managed Azure service that stores non-secret configuration settings in a secure,
centralized location. Stored values can be shared among multiple services and applications.
The service is simple to use and provides several benefits:
Flexible key/value representations and mappings
Tagging with Azure labels
Dedicated UI for management
Encryption of sensitive information
Querying and batch retrieval
Azure App Configuration maintains changes made to key-value settings for seven days. The point-in-time snapshot
feature enables you to reconstruct the history of a setting and even rollback for a failed deployment.
App Configuration automatically caches each setting to avoid excessive calls to the configuration store. The refresh
operation waits until the cached value of a setting expires to update that setting, even when its value changes in the
configuration store. The default cache expiration time is 30 seconds. You can override the expiration time.
App Configuration encrypts all configuration values in transit and at rest. Key names and labels are used as indexes
for retrieving configuration data and aren't encrypted.
Although App Configuration provides hardened security, Azure Key Vault is still the best place for storing
application secrets. Key Vault provides hardware-level encryption, granular access policies, and management
operations such as certificate rotation. You can create App Configuration values that reference secrets stored in a
Key Vault.
Azure Key Vault

Key Vault is a managed service for securely storing and accessing secrets. A secret is anything that you want to
tightly control access to, such as API keys, passwords, or certificates. A vault is a logical group of secrets.
Key Vault greatly reduces the chances that secrets may be accidentally leaked. When using Key Vault, application
developers no longer need to store security information in their application. This practice eliminates the need to
store this information inside your code. For example, an application may need to connect to a database. Instead of
storing the connection string in the app's code, you can store it securely in Key Vault.
Your applications can securely access the information they need by using URIs. These URIs allow the applications to
retrieve specific versions of a secret. There's no need to write custom code to protect any of the secret information
stored in Key Vault.
Access to Key Vault requires proper caller authentication and authorization. Typically, each cloud-native
microservice uses a ClientId/ClientSecret combination. It's important to keep these credentials outside source
control. A best practice is to set them in the application's environment. Direct access to Key Vault from AKS can be
achieved using Key Vault FlexVolume.
Configuration in eShop
The eShopOnContainers application includes local application settings files with each microservice. These files are
checked into source control, but don't include production secrets such as connection strings or API keys. In
production, individual settings may be overwritten with per-service environment variables. Injecting secrets in
environment variables is a common practice for hosted applications, but doesn't provide a central configuration
store. To support centralized management of configuration settings, each microservice includes a setting to toggle
between its use of local settings or Azure Key Vault settings.
References
The eShopOnContainers Architecture
Orchestrating microservices and multi-container applications for high scalability and availability
Azure API Management
Azure SQL Database Overview
Azure Cache for Redis
Azure Cosmos DB's API for MongoDB
Azure Service Bus
Azure Monitor overview
eShopOnContainers: Create Kubernetes cluster in AKS
eShopOnContainers: Azure Dev Spaces
Azure Dev Spaces
PR EVIO U S NE XT
Scaling cloud-native applications
One of the most-often touted advantages of moving to a cloud hosting environment is scalability. Scalability, or the
ability for an application to accept additional user load without compromising performance for each user. It's most
often achieved by breaking up an application into small pieces that can each be given whatever resources they
require. Cloud vendors enable massive scalability anytime and anywhere in the world.
In this chapter, we discuss technologies that enable cloud-native applications to scale to meet user demand. These
technologies include:
Containers
Orchestrators
Serverless computing
PR EVIO U S NE XT
Leveraging containers and orchestrators
Containers and orchestrators are designed to solve problems common to monolithic deployment approaches.
Challenges with monolithic deployments

Traditionally, most applications have been deployed as a single unit. Such applications are referred to as a
monolith. This general approach of deploying applications as single units even if they're composed of multiple
modules or assemblies is known as monolithic architecture, as shown in Figure 3-1.
Figure 3-1 . Monolithic architecture.

Although they have the benefit of simplicity, monolithic architectures face a number of challenges:
Deployment
Monolithic applications require a full deployment of the entire application, even if only a small change has been
made. Full deployments can be expensive and error prone. Additionally, they require a restart of the application,
which temporarily impacts unavailability.
Scaling
A monolithic application is hosted entirely on a single machine instance, often requiring high-capability hardware.
If any part of the monolith requires scaling, another copy of the entire application must be deployed to another
machine. With a monolith, you can't scale application components individually - it's all or nothing. Scaling
components that don't require scaling results in inefficient and costly resource usage.
Environment
Monolithic applications are typically deployed to a hosting environment with a pre-installed operating system,
runtime, and library dependencies. This environment may not match that upon which the application was
developed or tested. Inconsistencies across application environments are a common source of problems for
monolithic deployments.
Coupling
A monolithic application is likely to experience high coupling across its functional components. Without hard
boundaries, system changes often result in unintended and costly side effects. New features/fixes become tricky,
time-consuming, and expensive to implement. Updates require extensive testing. Coupling also makes it difficult to
refactor components or swap in alternative implementations. Even when constructed with a strict separation of
concerns, architectural erosion sets in as the monolithic code base deteriorates with never-ending "special cases."
Platform lock-in
A monolithic application is constructed with a single technology stack. While offering uniformity, this commitment
can become a barrier to innovation. New features and components will be built using the application's current
stack - even when more modern technologies may be a better choice. A longer-term risk is your technology stack
becoming outdated and obsolete. Rearchitecting an entire application to a new, more modern platform is at best
expensive and risky.
What are the benefits of containers and orchestrators?

We introduced containers in Chapter 1. We highlighted how the Cloud Native Computing Foundation (CNCF)
ranks containerization as the first step in their Cloud-Native Trail Map - guidance for enterprises beginning their
cloud-native journey. In this section, we discuss the benefits of containers.
Docker is the most popular container management platform. It works with containers on both Linux or Windows.
Containers provide separate but reproducible application environments that run the same way on any system. This
aspect makes them perfect for developing and hosting cloud-native services. Containers are isolated from one
another. Two containers on the same host hardware can have different versions of software, without causing
conflicts.
Containers are defined by simple text-based files that become project artifacts and are checked into source control.
While full servers and virtual machines require manual effort to update, containers are easily version-controlled.
Apps built to run in containers can be developed, tested, and deployed using automated tools as part of a build
pipeline.
Containers are immutable. Once you define a container, you can recreate and run it exactly the same way. This
immutability lends itself to component-based design. If some parts of an application evolve differently than others,
why redeploy the entire app when you can just deploy the parts that change most frequently? Different features
and cross-cutting concerns of an app can be broken up into separate units. Figure 3-2 shows how a monolithic app
can take advantage of containers and microservices by delegating certain features or functionality. The remaining
functionality in the app itself has also been containerized.
Figure 3-2 . Decomposing a monolithic app to embrace microservices.

Each cloud-native service is built and deployed in a separate container. Each can update as needed. Individual
services can be hosted on nodes with resources appropriate to each service. The environment each service runs in
is immutable, shared across dev, test, and production environments, and easily versioned. Coupling between
different areas of the application occurs explicitly as calls or messages between services, not compile-time
dependencies within the monolith. You can also choose the technology that best suites a given capability without
requiring changes to the rest of the app.
Containerized services require automated management. It wouldn't be feasible to manually administer a large set
of independently deployed containers. For example, consider the following tasks:
How will container instances be provisioned across a cluster of many machines?
Once deployed, how will containers discover and communicate with each other?
How can containers scale in or out on-demand?
How do you monitor the health of each container?
How do you protect a container against hardware and software failures?
How do upgrade containers for a live application with zero downtime?
Container orchestrators address and automate these and other concerns.
In the cloud-native eco-system, Kubernetes has become the de facto container orchestrator. It's an open-source
platform managed by the Cloud Native Computing Foundation (CNCF). Kubernetes automates the deployment,
scaling, and operational concerns of containerized workloads across a machine cluster. However, installing and
managing Kubernetes is notoriously complex.
A much better approach is to leverage Kubernetes as a managed service from a cloud vendor. The Azure cloud
features a fully managed Kubernetes platform entitled Azure Kubernetes Service (AKS). AKS abstracts the
complexity and operational overhead of managing Kubernetes. You consume Kubernetes as a cloud service;
Microsoft takes responsibility for managing and supporting it. AKS also tightly integrates with other Azure services
and dev tools.
AKS is a cluster-based technology. A pool of federated virtual machines, or nodes, is deployed to the Azure cloud.
Together they form a highly available environment, or cluster. The cluster appears as a seamless, single entity to
your cloud-native application. Under the hood, AKS deploys your containerized services across these nodes
following a predefined strategy that evenly distributes the load.
What are the scaling benefits?

Services built on containers can leverage scaling benefits provided by orchestration tools like Kubernetes. By
design containers only know about themselves. Once you have multiple containers that need to work together, you
should organize them at a higher level. Organizing large numbers of containers and their shared dependencies,
such as network configuration, is where orchestration tools come in to save the day! Kubernetes creates an
abstraction layer over groups of containers and organizes them into pods. Pods run on worker machines referred
to as nodes. This organized structure is referred to as a cluster. Figure 3-3 shows the different components of a
Kubernetes cluster.
Figure 3-3 . Kubernetes cluster components.
Scaling containerized workloads is a key feature of container orchestrators. AKS supports automatic scaling across
two dimensions: Container instances and compute nodes. Together they give AKS the ability to quickly and
efficiently respond to spikes in demand and add additional resources. We discuss scaling in AKS later in this
chapter.
Declarative versus imperative
Kubernetes supports both declarative and imperative configuration. The imperative approach involves running
various commands that tell Kubernetes what to do each step of the way. Run this image. Delete this pod. Expose
this port. With the declarative approach, you create a configuration file, called a manifest, to describe what you
want instead of what to do. Kubernetes reads the manifest and transforms your desired end state into actual end
state.
Imperative commands are great for learning and interactive experimentation. However, you'll want to declaratively
create Kubernetes manifest files to embrace an infrastructure as code approach, providing for reliable and
repeatable deployments. The manifest file becomes a project artifact and is used in your CI/CD pipeline for
automating Kubernetes deployments.
If you've already configured your cluster using imperative commands, you can export a declarative manifest by
using kubectl get svc SERVICENAME -o yaml > service.yaml . This command produces a manifest similar to one
shown below:
apiVersion: v1
kind: Service
metadata:
creationTimestamp: "2019-09-13T13:58:47Z"
labels:
component: apiserver
provider: kubernetes
name: kubernetes
namespace: default
resourceVersion: "153"
selfLink: /api/v1/namespaces/default/services/kubernetes
uid: 9b1fac62-d62e-11e9-8968-00155d38010d
spec:
clusterIP: 10.96.0.1
ports:
- name: https
port: 443
protocol: TCP
targetPort: 6443
sessionAffinity: None
type: ClusterIP
status:
loadBalancer: {}
When using declarative configuration, you can preview the changes that will be made before committing them by
using kubectl diff -f FOLDERNAME against the folder where your configuration files are located. Once you're sure
you want to apply the changes, run kubectl apply -f FOLDERNAME . Add -R to recursively process a folder hierarchy.
You can also use declarative configuration with other Kubernetes features, one of which being deployments.
Declarative deployments help manage releases, updates, and scaling. They instruct the Kubernetes deployment
controller on how to deploy new changes, scale out load, or roll back to a previous revision. If a cluster is unstable,
a declarative deployment will automatically return the cluster back to a desired state. For example, if a node should
crash, the deployment mechanism will redeploy a replacement to achieve your desired state
Using declarative configuration allows infrastructure to be represented as code that can be checked in and
versioned alongside the application code. It provides improved change control and better support for continuous
deployment using a build and deploy pipeline.
What scenarios are ideal for containers and orchestrators?

The following scenarios are ideal for using containers and orchestrators.
Applications requiring high uptime and scalability
Individual applications that have high uptime and scalability requirements are ideal candidates for cloud-native
architectures using microservices, containers, and orchestrators. They can be developed in containers, tested across
versioned environments, and deployed into production with zero downtime. The use of Kubernetes clusters
ensures such apps can also scale on demand and recover automatically from node failures.
Large numbers of applications
Organizations that deploy and maintain large numbers of applications benefit from containers and orchestrators.
The up front effort of setting up containerized environments and Kubernetes clusters is primarily a fixed cost.
Deploying, maintaining, and updating individual applications has a cost that varies with the number of applications.
Beyond a small number of applications, the complexity of maintaining custom applications manually exceeds the
cost of implementing a solution using containers and orchestrators.
When should you avoid using containers and orchestrators?

If you're unable to build your application following the Twelve-Factor App principles, you should consider avoiding
containers and orchestrators. In these cases, consider a VM-based hosting platform, or possibly some hybrid
system. With it, you can always spin off certain pieces of functionality into separate containers or even serverless
functions.
Development resources
This section shows a short list of development resources that may help you get started using containers and
orchestrators for your next application. If you're looking for guidance on how to design your cloud-native
microservices architecture app, read this book's companion, .NET Microservices: Architecture for Containerized
.NET Applications.
Local Kubernetes Development
Kubernetes deployments provide great value in production environments, but can also run locally on your
development machine. While you may work on individual microservices independently, there may be times when
you'll need to run the entire system locally - just as it will run when deployed to production. There are several tools
that can help: Minikube and Docker Desktop. Visual Studio also provides tooling for Docker development.
Minikube
What is Minikube? The Minikube project says "Minikube implements a local Kubernetes cluster on macOS, Linux,
and Windows." Its primary goals are "to be the best tool for local Kubernetes application development and to
support all Kubernetes features that fit." Installing Minikube is separate from Docker, but Minikube supports
different hypervisors than Docker Desktop supports. The following Kubernetes features are currently supported by
Minikube:
DNS
NodePorts
ConfigMaps and secrets
Dashboards
Container runtimes: Docker, rkt, CRI-O, and containerd
Enabling Container Network Interface (CNI)
Ingress
After installing Minikube, you can quickly start using it by running the minikube start command, which
downloads an image and start the local Kubernetes cluster. Once the cluster is started, you interact with it using the
standard Kubernetes kubectl commands.
Docker Desktop
You can also work with Kubernetes directly from Docker Desktop on Windows. It is your only option if you're using
Windows Containers, and is a great choice for non-Windows containers as well. Figure 3-4 shows how to enable
local Kubernetes support when running Docker Desktop.
Figure 3-4 . Configuring Kubernetes in Docker Desktop.
Docker Desktop is the most popular tool for configuring and running containerized apps locally. When you work
with Docker Desktop, you can develop locally against the exact same set of Docker container images that you'll
deploy to production. Docker Desktop is designed to "build, test, and ship" containerized apps locally. It supports
both Linux and Windows containers. Once you push your images to an image registry, like Azure Container
Registry or Docker Hub, AKS can pull and deploy them to production.
Visual Studio Docker Tooling
Visual Studio supports Docker development for web-based applications. When you create a new ASP.NET Core
application, you have an option to configure it with Docker support, as shown in Figure 3-5.
Figure 3-5 . Visual Studio Enable Docker Support
When this option is selected, the project is created with a Dockerfile in its root, which can be used to build and
host the app in a Docker container. An example Dockerfile is shown in Figure 3-6.git
FROM mcr.microsoft.com/dotnet/core/aspnet:3.1-buster-slim AS base

WORKDIR /app
EXPOSE 80
EXPOSE 443
FROM mcr.microsoft.com/dotnet/core/sdk:3.1-buster AS build

WORKDIR /src
COPY ["eShopWeb/eShopWeb.csproj", "eShopWeb/"]
RUN dotnet restore "eShopWeb/eShopWeb.csproj"
COPY . .
WORKDIR "/src/eShopWeb"
RUN dotnet build "eShopWeb.csproj" -c Release -o /app/build
FROM build AS publish

RUN dotnet publish "eShopWeb.csproj" -c Release -o /app/publish
FROM base AS final

WORKDIR /app
COPY --from=publish /app/publish .
ENTRYPOINT ["dotnet", "eShopWeb.dll"]
Figure 3-6 . Visual Studio generated Dockerfile

The default behavior when the app runs is configured to use Docker as well. Figure 3-7 shows the different run
options available from a new ASP.NET Core project created with Docker support added.
Figure 3-7 . Visual Studio Docker Run Options
In addition to local development, Azure Dev Spaces provides a convenient way for multiple developers to work
with their own Kubernetes configurations within Azure. As you can see in Figure 3-7, you can also run the
application in Azure Dev Spaces.
Also, at any time you can add Docker support to an existing ASP.NET Core application. From the Visual Studio
Solution Explorer, right click on the project and Add > Docker Suppor t , as shown in Figure 3-8.
Figure 3-8 . Adding Docker support to Visual Studio

You can also add Container Orchestration Support, also shown in Figure 3-8. By default, the orchestrator uses
Kubernetes and Helm. Once you've chosen the orchestrator, a azds.yaml file is added to the project root and a
charts folder is added containing the Helm charts used to configure and deploy the application to Kubernetes.
Figure 3-9 shows the resulting files in a new project.
Figure 3-9 . Adding orchestration support to Visual Studio

Visual Studio Code Docker Tooling
There are a number of extensions available for Visual Studio Code that support Docker development.
Microsoft provides the Docker for Visual Studio Code extension. This extension simplifies the process of adding
container support to applications. It scaffolds required files, builds Docker images, and enables you to debug your
app inside a container. The extension features a visual explorer that makes it easy to take actions on containers and
images such as start, stop, inspect, remove, and more. The extension also supports Docker Compose enabling you
to manage multiple running containers as a single unit.
PR EVIO U S NE XT
Leveraging serverless functions
In the spectrum from managing physical machines to leveraging cloud capabilities, serverless lives at the extreme
end. Your only responsibility is your code, and you only pay when your code runs. Azure Functions provides a way
to build serverless capabilities into your cloud-native applications.
What is serverless?
Serverless is a relatively new service model of cloud computing. It doesn't mean that servers are optional - your
code still runs on a server somewhere. The distinction is that the application team no longer concerns itself with
managing server infrastructure. Instead, the cloud vendor own this responsibility. The development team increases
its productivity by delivering business solutions to customers, not plumbing.
Serverless computing uses event-triggered stateless containers to host your services. They can scale out and in to
meet demand as-needed. Serverless platforms like Azure Functions have tight integration with other Azure
services like queues, events, and storage.
What challenges are solved by serverless?

Serverless platforms address many time-consuming and expensive concerns:
Purchasing machines and software licenses
Housing, securing, configuring, and maintaining the machines and their networking, power, and A/C
requirements
Patching and upgrading operating systems and software
Configuring web servers or machine services to host application software
Configuring application software within its platform
Many companies allocate large budgets to support hardware infrastructure concerns. Moving to the cloud can help
reduce these costs; shifting applications to serverless can help eliminate them.
What is the difference between a microservice and a serverless

function?
Typically, a microservice encapsulates a business capability, such as a shopping cart for an online eCommerce site.
It exposes multiple operations that enable a user to manage their shopping experience. A function, however, is a
small, lightweight block of code that executes a single-purpose operation in response to an event. Microservices
are typically constructed to respond to requests, often from an interface. Requests can be HTTP Rest- or gRPC-
based. Serverless services respond to events. Its event-driven architecture is ideal for processing short-running,
background tasks.
What scenarios are appropriate for serverless?

Serverless exposes individual short-running functions that are invoked in response to a trigger. This makes them
ideal for processing background tasks.
An application might need to send an email as a step in a workflow. Instead of sending the notification as part of a
microservice request, place the message details onto a queue. An Azure Function can dequeue the message and
asynchronously send the email. Doing so could improve the performance and scalability of the microservice.
Queue-based load leveling can be implemented to avoid bottlenecks related to sending the emails. Additionally,
this stand-alone service could be reused as a utility across many different applications.
Asynchronous messaging from queues and topics is a common pattern to trigger serverless functions. However,
Azure Functions can be triggered by other events, such as changes to Azure Blob Storage. A service that supports
image uploads could have an Azure Function responsible for optimizing the image size. The function could be
triggered directly by inserts into Azure Blob Storage, keeping complexity out of the microservice operations.
Many services have long-running processes as part of their workflows. Often these tasks are done as part of the
user's interaction with the application. These tasks can force the user to wait, negatively impacting their experience.
Serverless computing provides a great way to move slower tasks outside of the user interaction loop. These tasks
can scale with demand without requiring the entire application to scale.
When should you avoid serverless?

Serverless solutions provision and scale on demand. When a new instance is invoked, cold starts are a common
issue. A cold start is the period of time it takes to provision this instance. Normally, this delay might be a few
seconds, but can be longer depending on various factors. Once provisioned, a single instance is kept alive as long
as it receives periodic requests. But, if a service is called less frequently, Azure may remove it from memory and
require a cold start when reinvoked. Cold starts are also required when a function scales out to a new instance.
Figure 3-10 shows a cold-start pattern. Note the extra steps required when the app is cold.
Figure 3-10 . Cold start versus warm start.

To avoid cold starts entirely, you might switch from a consumption plan to a dedicated plan. You can also configure
one or more pre-warmed instances with the premium plan upgrade. In these cases, when you need to add another
instance, it's already up and ready to go. These options can help mitigate the cold start issue associated with
serverless computing.
Cloud providers bill for serverless based on compute execution time and consumed memory. Long running
operations or high memory consumption workloads aren't always the best candidates for serverless. Serverless
functions favor small chunks of work that can complete quickly. Most serverless platforms require individual
functions to complete within a few minutes. Azure Functions defaults to a 5-minute time-out duration, which can
be configured up to 10 minutes. The Azure Functions premium plan can mitigate this issue as well, defaulting time-
outs to 30 minutes with an unbounded higher limit that can be configured. Compute time isn't calendar time. More
advanced functions using the Azure Durable Functions framework may pause execution over a course of several
days. The billing is based on actual execution time - when the function wakes up and resumes processing.
Finally, leveraging Azure Functions for application tasks adds complexity. It's wise to first architect your application
with a modular, loosely coupled design. Then, identify if there are benefits serverless would offer that justify the
additional complexity.
PR EVIO U S NE XT
Combining containers and serverless approaches
Cloud-native applications typically implement services leveraging containers and orchestration. There are often
opportunities to expose some of the application's services as Azure Functions. However, with a cloud-native app
deployed to Kubernetes, it would be nice to leverage Azure Functions within this same toolset. Fortunately, you can
wrap Azure Functions inside Docker containers and deploy them using the same processes and tools as the rest of
your Kubernetes-based app.
When does it make sense to use containers with serverless?

Your Azure Function has no knowledge of the platform on which it's deployed. For some scenarios, you may have
specific requirements and need to customize the environment on which your function code will run. You'll need a
custom image that supports dependencies or a configuration not supported by the default image. In these cases, it
makes sense to deploy your function in a custom Docker container.
When should you avoid using containers with Azure Functions?

If you want to use consumption billing, you won't be able to run your function in a container. What's more, if you
deploy your function to a Kubernetes cluster, you'll no longer benefit from the built-in scaling provided by Azure
Functions. You'll need to use Kubernetes' scaling features, described earlier in this chapter.
How to combine serverless and Docker containers

To wrap an Azure Function in a Docker container, install the Azure Functions Core Tools and then run the following
command:
func init ProjectName --worker-runtime dotnet --docker
When the project is created, it will include a Dockerfile and the worker runtime configured to dotnet . Now, you
can create and test your function locally. Build and run it using the docker build and docker run commands. For
detailed steps to get started building Azure Functions with Docker support, see the Create a function on Linux
using a custom image tutorial.
How to combine serverless and Kubernetes with KEDA

In this chapter, you've seen that the Azure Functions' platform automatically scales out to meet demand. When
deploying containerized functions to AKS, however, you lose the built-in scaling functionality. To the rescue comes
Kubernetes-based Event Driven (KEDA). It enables fine-grained autoscaling for event-driven Kubernetes workloads,
including containerized functions.
KEDA provides event-driven scaling functionality to the Functions' runtime in a Docker container. KEDA can scale
from zero instances (when no events are occurring) out to n instances , based on load. It enables autoscaling by
exposing custom metrics to the Kubernetes autoscaler (Horizontal Pod Autoscaler). Using Functions containers with
KEDA makes it possible to replicate serverless function capabilities in any Kubernetes cluster.
It is worth noting that the KEDA project is now managed by the Cloud Native Computing Foundation (CNCF).
PR EVIO U S NE XT
Deploying containers in Azure
We've discussed containers in this chapter and in chapter 1. We've seen that containers provide many benefits to
cloud-native applications, including portability. In the Azure cloud, you can deploy the same containerized services
across staging and production environments. Azure provides several options for hosting these containerized
workloads:
Azure Kubernetes Services (AKS)
Azure Container Instance (ACI)
Azure Web Apps for Containers
Azure Container Registry

When containerizing a microservice, you first a build container "image." The image is a binary representation of
the service code, dependencies, and runtime. While you can manually create an image using the Docker Build
command from the Docker API, a better approach is to create it as part of an automated build process.
Once created, container images are stored in container registries. They enable you to build, store, and manage
container images. There are many registries available, both public and private. Azure Container Registry (ACR) is a
fully managed container registry service in the Azure cloud. It persists your images inside the Azure network,
reducing the time to deploy them to Azure container hosts. You can also secure them using the same security and
identity procedures that you use for other Azure resources.
You create an Azure Container Registry using the Azure portal, Azure CLI, or PowerShell tools. Creating a registry in
Azure is simple. It requires an Azure subscription, resource group, and a unique name. Figure 3-11 shows the basic
options for creating a registry, which will be hosted at registryname.azurecr.io .
Figure 3-11 . Create container registry

Once you've created the registry, you'll need to authenticate with it before you can use it. Typically, you'll log into
the registry using the Azure CLI command:
az acr login --name *registryname*
Once authenticated, you can use docker commands to push container images to it. Before you can do so, however,
you must tag your image with the fully qualified name (URL) of your ACR login server. It will have the format
registryname.azurecr.io.
docker tag mycontainer myregistry.azurecr.io/mycontainer:v1
After you've tagged the image, you use the docker push command to push the image to your ACR instance.
docker push myregistry.azurecr.io/mycontainer:v1
After you push an image to the registry, it's a good idea to remove the image from your local Docker environment,
using this command:
docker rmi myregistry.azurecr.io/mycontainer:v1
As a best practice, developers shouldn't manually push images to a container registry. Instead, a build pipeline
defined in a tool like GitHub or Azure DevOps should be responsible for this process. Learn more in the Cloud-
Native DevOps chapter.
ACR Tasks
ACR Tasks is a set of features available from the Azure Container Registry. It extends your inner-loop development
cycle by building and managing container images in the Azure cloud. Instead of invoking a docker build and
docker push locally on your development machine, they're automatically handled by ACR Tasks in the cloud.
The following AZ CLI command both builds a container image and pushes it to ACR:
# create a container registry

az acr create --resource-group myResourceGroup --name myContainerRegistry008 --sku Basic
# build container image in ACR and push it into your container regsitry
az acr build --image sample/hello-world:v1 --registry myContainerRegistry008 --file Dockerfile .
As you can see from the previous command block, there's no need to install Docker Desktop on your development
machine. Additionally, you can configure ACR Task triggers to rebuild containers images on both source code and
base image updates.
Azure Kubernetes Service

We discussed Azure Kubernetes Service (AKS) at length in this chapter. We've seen that it's the de facto container
orchestrator managing containerized cloud-native applications.
Once you deploy an image to a registry, such as ACR, you can configure AKS to automatically pull and deploy it.
With a CI/CD pipeline in place, you might configure a canary release strategy to minimize the risk involved when
rapidly deploying updates. The new version of the app is initially configured in production with no traffic routed to
it. Then, the system will route a small percentage of users to the newly deployed version. As the team gains
confidence in the new version, it can roll out more instances and retire the old. AKS easily supports this style of
deployment.
As with most resources in Azure, you can create an Azure Kubernetes Service cluster using the portal, command-
line, or automation tools like Helm or Terraform. To get started with a new cluster, you need to provide the
following information:
Azure subscription
Resource group
Kubernetes cluster name
Region
Kubernetes version
DNS name prefix
Node size
Node count
This information is sufficient to get started. As part of the creation process in the Azure portal, you can also
configure options for the following features of your cluster:
Scale
Authentication
Networking
Monitoring
Tags
This quickstart walks through deploying an AKS cluster using the Azure portal.
Azure Dev Spaces

Cloud-native applications can quickly grow large and complex, requiring significant compute resources to run. In
these scenarios, the entire application can't be hosted on a development machine (especially a laptop). Azure Dev
Spaces is designed to address this problem using AKS. It enables developers to work with a local version of their
services while hosting the rest of the application in an AKS development cluster.
Developers share a running (development) instance in an AKS cluster that contains the entire containerized
application. But they use personal spaces set up on their machine to locally develop their services. When ready,
they test from end-to-end in the AKS cluster - without replicating dependencies. Azure Dev Spaces merges code
from the local machine with services in AKS. Developers can rapidly iterate and debug code directly in Kubernetes
using Visual Studio or Visual Studio Code.
To understand the value of Azure Dev Spaces, let me share this quotation from Gabe Monroy, PM Lead of
Containers at Microsoft Azure:
"Imagine you're a new employee trying to fix a bug in a complex microservices application consisting of
dozens of components, each with their own configuration and backing services. To get started, you must
configure your local development environment so that it can mimic production including setting up your IDE,
building tool chain, containerized service dependencies, a local Kubernetes environment, mocks for backing
services, and more. With all the time involved setting up your development environment, fixing that first bug
could take days. Or you could use Dev Spaces and AKS."
The process for working with Azure Dev Spaces involves the following steps:
1. Create the dev space.
2. Configure the root dev space.
3. Configure a child dev space (for your own version of the system).
4. Connect to the dev space.
All of these steps can be performed using the Azure CLI and new azds command-line tools. For example, to create
a new Azure Dev Space for a given Kubernetes cluster, you would use a command like this one:
az aks use-dev-spaces -g my-aks-resource-group -n MyAKSCluster
Next, you can use the azds prep command to generate the necessary Docker and Helm chart assets for running
the application. Then you run your code in AKS using azds up . The first time you run this command, the Helm
chart will be installed. The containers will be built and deployed according to your instructions. This task may take a
few minutes the first time it's run. However, after you make changes, you can connect to your own child dev space
using azds space select and then deploy and debug your updates in your isolated child dev space. Once you have
your dev space up and running, you can send updates to it by reissuing the azds up command or you can use
built-in tooling in Visual Studio or Visual Studio Code. With VS Code, you use the command palette to connect to
your dev space. Figure 3-12 shows how to launch your web application using Azure Dev Spaces in Visual Studio.
Figure 3-12 . Connect to Azure Dev Spaces in Visual Studio
PR EVIO U S NE XT
Scaling containers and serverless applications
There are two ways to scale an application: up or out. The former refers to adding capacity to a single resource,
while the latter refers to adding more resources to increase capacity.
The simple solution: scaling up

Upgrading an existing host server with increased CPU, memory, disk I/O speed, and network I/O speed is known as
scaling up. Scaling up a cloud-native application involves choosing more capable resources from the cloud vendor.
For example, you can a new node pool with larger VMs in your Kubernetes cluster. Then, migrate your
containerized services to the new pool.
Serverless apps scale up by choosing the premium Functions plan or premium instance sizes from a dedicated app
service plan.
Scaling out cloud-native apps

Cloud-native applications often experience large fluctuations in demand and require scale on a moment's notice.
They favor scaling out. Scaling out is done horizontally by adding additional machines (called nodes) or application
instances to an existing cluster. In Kubernetes, you can scale manually by adjusting configuration settings for the
app (for example, scaling a node pool), or through autoscaling.
AKS clusters can autoscale in one of two ways:
First, the Horizontal Pod Autoscaler monitors resource demand and automatically scales your POD replicas to meet
it. When traffic increases, additional replicas are automatically provisioned to scale out your services. Likewise,
when demand decreases, they're removed to scale-in your services. You define the metric on which to scale, for
example, CPU usage. You can also specify the minimum and maximum number of replicas to run. AKS monitors
that metric and scales accordingly.
Next, the AKS Cluster Autoscaler feature enables you to automatically scale compute nodes across a Kubernetes
cluster to meet demand. With it, you can automatically add new VMs to the underlying Azure Virtual Machine Scale
Set whenever more compute capacity of is required. It also removes nodes when no longer required.
Figure 3-13 shows the relationship between these two scaling services.
Figure 3-13 . Scaling out an App Service plan.

Working together, both ensure an optimal number of container instances and compute nodes to support
fluctuating demand. The horizontal pod autoscaler optimizes the number of pods required. The cluster autoscaler
optimizes the number of nodes required.
Scaling Azure Functions
Azure Functions automatically scale out upon demand. Server resources are dynamically allocated and removed
based on the number of triggered events. You're only charged for compute resources consumed when your
functions run. Billing is based upon the number of executions, execution time, and memory used.
While the default consumption plan provides an economical and scalable solution for most apps, the premium
option allows developers flexibility for custom Azure Functions requirements. Upgrading to the premium plan
provides control over instance sizes, pre-warmed instances (to avoid cold start delays), and dedicated VMs.
PR EVIO U S NE XT
Other container deployment options
Aside from Azure Kubernetes Service (AKS), you can also deploy containers to Azure App Service for Containers
and Azure Container Instances.
When does it make sense to deploy to App Service for Containers?

Simple production applications that don't require orchestration are well suited to Azure App Service for Containers.
How to deploy to App Service for Containers

To deploy to Azure App Service for Containers, you'll need an Azure Container Registry (ACR) instance and
credentials to access it. Push your container image to the ACR repository so it can pulled into your Azure App
Service. Once complete, you can configure the app for Continuous Deployment. Doing so will automatically deploy
updates whenever the image changes in ACR.
When does it make sense to deploy to Azure Container Instances?

Azure Container Instances (ACI) enables you to run Docker containers in a managed, serverless cloud environment,
without having to set up virtual machines or clusters. It's a great solution for short-running workloads that can run
in an isolated container. Consider ACI for simple services, testing scenarios, task automation, and build jobs. ACI
spins-up a container instance, performs the task, and then spins it down.
How to deploy an app to Azure Container Instances

To deploy to Azure Container Instances (ACI), you need an Azure Container Registry (ACR) and credentials for
accessing it. Once you push your container image to the repository, it's available to pull into ACI. You can work with
ACI using the Azure portal or command-line interface. ACR provides tight integration with ACI. Figure 3-14 shows
how to push an individual container image to ACR.
Figure 3-14 . Azure Container Registry Run Instance
Creating an instance in ACI can be done quickly. Specify the image registry, Azure resource group information, the
amount of memory to allocate, and the port on which to listen. This quickstart shows how to deploy a container
instance to ACI using the Azure portal.
Once the deployment completes, find the newly deployed container's IP address and communicate with it over the
port you specified.
Azure Container Instances offers the fastest way to run simple container workloads in Azure. You don't need to
configure an app service, orchestrator, or virtual machine. For scenarios where you require full container
orchestration, service discovery, automatic scaling, or coordinated upgrades, we recommend Azure Kubernetes
Service (AKS).
References
What is Kubernetes?
Installing Kubernetes with Minikube
MiniKube vs Docker Desktop
Visual Studio Tools for Docker
Understanding serverless cold start
Pre-warmed Azure Functions instances
Create a function on Linux using a custom image
Run Azure Functions in a Docker Container
Create a function on Linux using a custom image
Azure Functions with Kubernetes Event Driven Autoscaling
Canary Release
Azure Dev Spaces with VS Code
Azure Dev Spaces with Visual Studio
AKS Multiple Node Pools
AKS Cluster Autoscaler
Tutorial: Scale applications in AKS
Azure Functions scale and hosting
Azure Container Instances Docs
Deploy Container Instance from ACR
PR EVIO U S NE XT
Cloud-native communication patterns
When constructing a cloud-native system, communication becomes a significant design decision. How does a
front-end client application communicate with a back-end microservice? How do back-end microservices
communicate with each other? What are the principles, patterns, and best practices to consider when
implementing communication in cloud-native applications?
Communication considerations
In a monolithic application, communication is straightforward. The code modules execute together in the same
executable space (process) on a server. This approach can have performance advantages as everything runs
together in shared memory, but results in tightly coupled code that becomes difficult to maintain, evolve, and scale.
Cloud-native systems implement a microservice-based architecture with many small, independent microservices.
Each microservice executes in a separate process and typically runs inside a container that is deployed to a cluster.
A cluster groups a pool of virtual machines together to form a highly available environment. They're managed with
an orchestration tool, which is responsible for deploying and managing the containerized microservices. Figure 4-1
shows a Kubernetes cluster deployed into the Azure cloud with the fully managed Azure Kubernetes Services.
Figure 4-1 . A Kubernetes cluster in Azure

Across the cluster, microservices communicate with each other through APIs and messaging technologies.
While they provide many benefits, microservices are no free lunch. Local in-process method calls between
components are now replaced with network calls. Each microservice must communicate over a network protocol,
which adds complexity to your system:
Network congestion, latency, and transient faults are a constant concern.
Resiliency (that is, retrying failed requests) is essential.
Some calls must be idempotent as to keep consistent state.
Each microservice must authenticate and authorize calls.
Each message must be serialized and then deserialized - which can be expensive.
Message encryption/decryption becomes important.
The book .NET Microservices: Architecture for Containerized .NET Applications, available for free from Microsoft,
provides an in-depth coverage of communication patterns for microservice applications. In this chapter, we provide
a high-level overview of these patterns along with implementation options available in the Azure cloud.
In this chapter, we'll first address communication between front-end applications and back-end microservices. We'll
then look at back-end microservices communicate with each other. We'll explore the up and gRPC communication
technology. Finally, we'll look new innovative communication patterns using service mesh technology. We'll also
see how the Azure cloud provides different kinds of backing services to support cloud-native communication.
PR EVIO U S NE XT
Front-end client communication
In a cloud-native system, front-end clients (mobile, web, and desktop applications) require a communication
channel to interact with independent back-end microservices.
What are the options?
To keep things simple, a front-end client could directly communicate with the back-end microservices, shown in
Figure 4-2.
Figure 4-2. Direct client to service communication

With this approach, each microservice has a public endpoint that is accessible by front-end clients. In a production
environment, you'd place a load balancer in front of the microservices, routing traffic proportionately.
While simple to implement, direct client communication would be acceptable only for simple microservice
applications. This pattern tightly couples front-end clients to core back-end services, opening the door for a
number of problems, including:
Client susceptibility to back-end service refactoring.
A wider attack surface as core back-end services are directly exposed.
Duplication of cross-cutting concerns across each microservice.
Overly complex client code - clients must keep track of multiple endpoints and handle failures in a resilient way.
Instead, a widely accepted cloud design pattern is to implement an API Gateway Service between the front-end
applications and back-end services. The pattern is shown in Figure 4-3.
Figure 4-3. API gateway pattern
In the previous figure, note how the API Gateway service abstracts the back-end core microservices. Implemented
as a web API, it acts as a reverse proxy, routing incoming traffic to the internal microservices.
The gateway insulates the client from internal service partitioning and refactoring. If you change a back-end
service, you accommodate for it in the gateway without breaking the client. It's also your first line of defense for
cross-cutting concerns, such as identity, caching, resiliency, metering, and throttling. Many of these cross-cutting
concerns can be off-loaded from the back-end core services to the gateway, simplifying the back-end services.
Care must be taken to keep the API Gateway simple and fast. Typically, business logic is kept out of the gateway. A
complex gateway risks becoming a bottleneck and eventually a monolith itself. Larger systems often expose
multiple API Gateways segmented by client type (mobile, web, desktop) or back-end functionality. The Backend for
Frontends pattern provides direction for implementing multiple gateways. The pattern is shown in Figure 4-4.
Figure 4-4. Backend for frontend pattern

Note in the previous figure how incoming traffic is sent to a specific API gateway - based upon client type: web,
mobile, or desktop app. This approach makes sense as the capabilities of each device differ significantly across
form factor, performance, and display limitations. Typically mobile applications expose less functionality than a
browser or desktop applications. Each gateway can be optimized to match the capabilities and functionality of the
corresponding device.
To start, you could build your own API Gateway service. A quick search of GitHub will provide many examples.
However, there are several frameworks and commercial gateway products available.
Ocelot Gateway
For simple .NET cloud-native applications, you might consider the Ocelot Gateway. Ocelot is an Open Source API
Gateway created for .NET microservices that require a unified point of entry into their system. It's lightweight, fast,
scalable.
Like any API Gateway, its primary functionality is to forward incoming HTTP requests to downstream services.
Additionally, it supports a wide variety of capabilities that are configurable in a .NET Core middleware pipeline. Its
feature set is presented in following table.
O C ELOT F EAT URES
Routing Authentication
Request Aggregation Authorization
Service Discovery (with Consul and Eureka) Throttling
Load Balancing Logging, Tracing
Caching Headers/Query String Transformation
Correlation Pass-Through Custom Middleware
Quality of Service Retry Policies
Each Ocelot gateway specifies the upstream and downstream addresses and configurable features in a JSON
configuration file. The client sends an HTTP request to the Ocelot gateway. Once received, Ocelot passes the
HttpRequest object through its pipeline manipulating it into the state specified by its configuration. At the end of
pipeline, Ocelot creates a new HTTPResponseObject and passes it to the downstream service. For the response,
Ocelot reverses the pipeline, sending the response back to client.
Ocelot is available as a NuGet package. It targets the NET Standard 2.0, making it compatible with both .NET Core
2.0+ and .NET Framework 4.6.1+ runtimes. Ocelot integrates with anything that speaks HTTP and runs on the
platforms which .NET Core supports: Linux, macOS, and Windows. Ocelot is extensible and supports many modern
platforms, including Docker containers, Azure Kubernetes Services, or other public clouds. Ocelot integrates with
open-source packages like Consul, GraphQL, and Netflix's Eureka.
Consider Ocelot for simple cloud-native applications that don't require the rich feature-set of a commercial API
gateway.
Azure Application Gateway

For simple gateway requirements, you may consider Azure Application Gateway. Available as an Azure PaaS
service, it includes basic gateway features such as URL routing, SSL termination, and a Web Application Firewall.
The service supports Layer-7 load balancing capabilities. With Layer 7, you can route requests based on the actual
content of an HTTP message, not just low-level TCP network packets.
Throughout this book, we evangelize hosting cloud-native systems in Kubernetes. A container orchestrator,
Kubernetes automates the deployment, scaling, and operational concerns of containerized workloads. Azure
Application Gateway can be configured as an API gateway for Azure Kubernetes Service cluster.
The Application Gateway Ingress Controller enables Azure Application Gateway to work directly with Azure
Kubernetes Service. Figure 4.5 shows the architecture.
Figure 4-5. Application Gateway Ingress Controller

Kubernetes includes a built-in feature that supports HTTP (Level 7) load balancing, called Ingress. Ingress defines a
set of rules for how microservice instances inside AKS can be exposed to the outside world. In the previous image,
the ingress controller interprets the ingress rules configured for the cluster and automatically configures the Azure
Application Gateway. Based on those rules, the Application Gateway routes traffic to microservices running inside
AKS. The ingress controller listens for changes to ingress rules and makes the appropriate changes to the Azure
Application Gateway.

For moderate to large-scale cloud-native systems, you may consider Azure API Management. It's a cloud-based
service that not only solves your API Gateway needs, but provides a full-featured developer and administrative
experience. API Management is shown in Figure 4-6.
Figure 4-6. Azure API Management
To start, API Management exposes a gateway server that allows controlled access to back-end services based upon
configurable rules and policies. These services can be in the Azure cloud, your on-prem data center, or other public
clouds. API keys and JWT tokens determine who can do what. All traffic is logged for analytical purposes.
For developers, API Management offers a developer portal that provides access to services, documentation, and
sample code for invoking them. Developers can use Swagger/Open API to inspect service endpoints and analyze
their usage. The service works across the major development platforms: .NET, Java, Golang, and more.
The publisher portal exposes a management dashboard where administrators expose APIs and manage their
behavior. Service access can be granted, service health monitored, and service telemetry gathered. Administrators
apply policies to each endpoint to affect behavior. Policies are pre-built statements that execute sequentially for
each service call. Policies are configured for an inbound call, outbound call, or invoked upon an error. Policies can
be applied at different service scopes as to enable deterministic ordering when combining policies. The product
ships with a large number of prebuilt policies.
Here are examples of how policies can affect the behavior of your cloud-native services:
Restrict service access.
Enforce authentication.
Throttle calls from a single source, if necessary.
Enable caching.
Block calls from specific IP addresses.
Control the flow of the service.
Convert requests from SOAP to REST or between different data formats, such as from XML to JSON.
Azure API Management can expose back-end services that are hosted anywhere – in the cloud or your data center.
For legacy services that you may expose in your cloud-native systems, it supports both REST and SOAP APIs. Even
other Azure services can be exposed through API Management. You could place a managed API on top of an Azure
backing service like Azure Service Bus or Azure Logic Apps. Azure API Management doesn't include built-in load-
balancing support and should be used in conjunction with a load-balancing service.
Azure API Management is available across four different tiers:
Developer
Basic
Standard
Premium
The Developer tier is meant for non-production workloads and evaluation. The other tiers offer progressively more
power, features, and higher service level agreements (SLAs). The Premium tier provides Azure Virtual Network and
multi-region support. All tiers have a fixed price per hour.
The Azure cloud also offers a serverless tier for Azure API Management. Referred to as the consumption pricing
tier, the service is a variant of API Management designed around the serverless computing model. Unlike the "pre-
allocated" pricing tiers previously shown, the consumption tier provides instant provisioning and pay-per-action
pricing.
It enables API Gateway features for the following use cases:
Microservices implemented using serverless technologies such as Azure Functions and Azure Logic Apps.
Azure backing service resources such as Service Bus queues and topics, Azure storage, and others.
Microservices where traffic has occasional large spikes but remains low the majority of the time.
The consumption tier uses the same underlying service API Management components, but employs an entirely
different architecture based on dynamically allocated resources. It aligns perfectly with the serverless computing
model:
No infrastructure to manage.
No idle capacity.
High-availability.
Automatic scaling.
Cost is based on actual usage.
The new consumption tier is a great choice for cloud-native systems that expose serverless resources as APIs.
Real-time communication
Real-time, or push, communication is another option for front-end applications that communicate with back-end
cloud-native systems over HTTP. Applications, such as financial-tickers, online education, gaming, and job-progress
updates, require instantaneous, real-time responses from the back-end. With normal HTTP communication, there's
no way for the client to know when new data is available. The client must continually poll or send requests to the
server. With real-time communication, the server can push new data to the client at any time.
Real-time systems are often characterized by high-frequency data flows and large numbers of concurrent client
connections. Manually implementing real-time connectivity can quickly become complex, requiring non-trivial
infrastructure to ensure scalability and reliable messaging to connected clients. You could find yourself managing
an instance of Azure Redis Cache and a set of load balancers configured with sticky sessions for client affinity.
Azure SignalR Service is a fully managed Azure service that simplifies real-time communication for your cloud-
native applications. Technical implementation details like capacity provisioning, scaling, and persistent connections
are abstracted away. They're handled for you with a 99.9% service-level agreement. You focus on application
features, not infrastructure plumbing.
Once enabled, a cloud-based HTTP service can push content updates directly to connected clients, including
browser, mobile and desktop applications. Clients are updated without the need to poll the server. Azure SignalR
abstracts the transport technologies that create real-time connectivity, including WebSockets, Server-Side Events,
and Long Polling. Developers focus on sending messages to all or specific subsets of connected clients.
Figure 4-7 shows a set of HTTP Clients connecting to a Cloud-native application with Azure SignalR enabled.
Figure 4-7. Azure SignalR
Another advantage of Azure SignalR Service comes with implementing Serverless cloud-native services. Perhaps
your code is executed on demand with Azure Functions triggers. This scenario can be tricky because your code
doesn't maintain long connections with clients. Azure SignalR Service can handle this situation since the service
already manages connections for you.
Azure SignalR Service closely integrates with other Azure services, such as Azure SQL Database, Service Bus, or
Redis Cache, opening up many possibilities for your cloud-native applications.
PR EVIO U S NE XT
Service-to-service communication
Moving from the front-end client, we now address back-end microservices communicate with each other.
When constructing a cloud-native application, you'll want to be sensitive to how back-end services communicate
with each other. Ideally, the less inter-service communication, the better. However, avoidance isn't always possible
as back-end services often rely on one another to complete an operation.
There are several widely accepted approaches to implementing cross-service communication. The type of
communication interaction will often determine the best approach.
Consider the following interaction types:
Query – when a calling microservice requires a response from a called microservice, such as, "Hey, give me
the buyer information for a given customer Id."
Command – when the calling microservice needs another microservice to execute an action but doesn't
require a response, such as, "Hey, just ship this order."
Event – when a microservice, called the publisher, raises an event that state has changed or an action has
occurred. Other microservices, called subscribers, who are interested, can react to the event appropriately.
The publisher and the subscribers aren't aware of each other.
Microservice systems typically use a combination of these interaction types when executing operations that require
cross-service interaction. Let's take a close look at each and how you might implement them.
Queries
Many times, one microservice might need to query another, requiring an immediate response to complete an
operation. A shopping basket microservice may need product information and a price to add an item to its basket.
There are a number of approaches for implementing query operations.
Request/Response Messaging
One option for implementing this scenario is for the calling back-end microservice to make direct HTTP requests to
the microservices it needs to query, shown in Figure 4-8.
Figure 4-8 . Direct HTTP communication
While direct HTTP calls between microservices are relatively simple to implement, care should be taken to
minimize this practice. To start, these calls are always synchronous and will block the operation until a result is
returned or the request times outs. What were once self-contained, independent services, able to evolve
independently and deploy frequently, now become coupled to each other. As coupling among microservices
increase, their architectural benefits diminish.
Executing an infrequent request that makes a single direct HTTP call to another microservice might be acceptable
for some systems. However, high-volume calls that invoke direct HTTP calls to multiple microservices aren't
advisable. They can increase latency and negatively impact the performance, scalability, and availability of your
system. Even worse, a long series of direct HTTP communication can lead to deep and complex chains of
synchronous microservices calls, shown in Figure 4-9:
Figure 4-9 . Chaining HTTP queries

You can certainly imagine the risk in the design shown in the previous image. What happens if Step #3 fails? Or
Step #8 fails? How do you recover? What if Step #6 is slow because the underlying service is busy? How do you
continue? Even if all works correctly, think of the latency this call would incur, which is the sum of the latency of
each step.
The large degree of coupling in the previous image suggests the services weren't optimally modeled. It would
behoove the team to revisit their design.
Materialized View pattern
A popular option for removing microservice coupling is the Materialized View pattern. With this pattern, a
microservice stores its own local, denormalized copy of data that's owned by other services. Instead of the
Shopping Basket microservice querying the Product Catalog and Pricing microservices, it maintains its own local
copy of that data. This pattern eliminates unnecessary coupling and improves reliability and response time. The
entire operation executes inside a single process. We explore this pattern and other data concerns in Chapter 5.
Service Aggregator Pattern
Another option for eliminating microservice-to-microservice coupling is an Aggregator microservice, shown in
purple in Figure 4-10.
Figure 4-10 . Aggregator microservice
The pattern isolates an operation that makes calls to multiple back-end microservices, centralizing its logic into a
specialized microservice. The purple checkout aggregator microservice in the previous figure orchestrates the
workflow for the Checkout operation. It includes calls to several back-end microservices in a sequenced order. Data
from the workflow is aggregated and returned to the caller. While it still implements direct HTTP calls, the
aggregator microservice reduces direct dependencies among back-end microservices.
Request/Reply Pattern
Another approach for decoupling synchronous HTTP messages is a Request-Reply Pattern, which uses queuing
communication. Communication using a queue is always a one-way channel, with a producer sending the message
and consumer receiving it. With this pattern, both a request queue and response queue are implemented, shown in
Figure 4-11.
Figure 4-11 . Request-reply pattern

Here, the message producer creates a query-based message that contains a unique correlation ID and places it into
a request queue. The consuming service dequeues the messages, processes it and places the response into the
response queue with the same correlation ID. The producer service dequeues the message, matches it with the
correlation ID and continues processing. We cover queues in detail in the next section.
Commands
Another type of communication interaction is a command. A microservice may need another microservice to
perform an action. The Ordering microservice may need the Shipping microservice to create a shipment for an
approved order. In Figure 4-12, one microservice, called a Producer, sends a message to another microservice, the
Consumer, commanding it to do something.
Figure 4-12 . Command interaction with a queue

Most often, the Producer doesn't require a response and can fire-and-forget the message. If a reply is needed, the
Consumer sends a separate message back to Producer on another channel. A command message is best sent
asynchronously with a message queue. supported by a lightweight message broker. In the previous diagram, note
how a queue separates and decouples both services.
A message queue is an intermediary construct through which a producer and consumer pass a message. Queues
implement an asynchronous, point-to-point messaging pattern. The Producer knows where a command needs to
be sent and routes appropriately. The queue guarantees that a message is processed by exactly one of the
consumer instances that are reading from the channel. In this scenario, either the producer or consumer service
can scale out without affecting the other. As well, technologies can be disparate on each side, meaning that we
might have a Java microservice calling a Golang microservice.
In chapter 1, we talked about backing services. Backing services are ancillary resources upon which cloud-native
systems depend. Message queues are backing services. The Azure cloud supports two types of message queues
that your cloud-native systems can consume to implement command messaging: Azure Storage Queues and Azure
Service Bus Queues.
Azure Storage Queues
Azure storage queues offer a simple queueing infrastructure that is fast, affordable, and backed by Azure storage
accounts.
Azure Storage Queues feature a REST-based queuing mechanism with reliable and persistent messaging. They
provide a minimal feature set, but are inexpensive and store millions of messages. Their capacity ranges up to 500
TB. A single message can be up to 64 KB in size.
You can access messages from anywhere in the world via authenticated calls using HTTP or HTTPS. Storage queues
can scale out to large numbers of concurrent clients to handle traffic spikes.
That said, there are limitations with the service:
Message order isn't guaranteed.
A message can only persist for seven days before it's automatically removed.
Support for state management, duplicate detection, or transactions isn't available.
Figure 4-13 shows the hierarchy of an Azure Storage Queue.
Figure 4-13 . Storage queue hierarchy

In the previous figure, note how storage queues store their messages in the underlying Azure Storage account.
For developers, Microsoft provides several client and server-side libraries for Storage queue processing. Most
major platforms are supported including .NET, Java, JavaScript, Ruby, Python, and Go. Developers should never
communicate directly with these libraries. Doing so will tightly couple your microservice code to the Azure Storage
Queue service. It's a better practice to insulate the implementation details of the API. Introduce an intermediation
layer, or intermediate API, that exposes generic operations and encapsulates the concrete library. This loose
coupling enables you to swap out one queuing service for another without having to make changes to the
mainline service code.
Azure Storage queues are an economical option to implement command messaging in your cloud-native
applications. Especially when a queue size will exceed 80 GB, or a simple feature set is acceptable. You only pay for
the storage of the messages; there are no fixed hourly charges.
Azure Service Bus Queues
For more complex messaging requirements, consider Azure Service Bus queues.
Sitting atop a robust message infrastructure, Azure Service Bus supports a brokered messaging model. Messages
are reliably stored in a broker (the queue) until received by the consumer. The queue guarantees First-In/First-Out
(FIFO) message delivery, respecting the order in which messages were added to the queue.
The size of a message can be much larger, up to 256 KB. Messages are persisted in the queue for an unlimited
period of time. Service Bus supports not only HTTP-based calls, but also provides full support for the AMQP
protocol. AMQP is an open-standard across vendors that supports a binary protocol and higher degrees of
reliability.
Service Bus provides a rich set of features, including transaction support and a duplicate detection feature. The
queue guarantees "at most once delivery" per message. It automatically discards a message that has already been
sent. If a producer is in doubt, it can resend the same message, and Service Bus guarantees that only one copy will
be processed. Duplicate detection frees you from having to build additional infrastructure plumbing.
Two more enterprise features are partitioning and sessions. A conventional Service Bus queue is handled by a
single message broker and stored in a single message store. But, Service Bus Partitioning spreads the queue across
multiple message brokers and message stores. The overall throughput is no longer limited by the performance of a
single message broker or messaging store. A temporary outage of a messaging store doesn't render a partitioned
queue unavailable.
Service Bus Sessions provide a way to group-related messages. Imagine a workflow scenario where messages
must be processed together and the operation completed at the end. To take advantage, sessions must be explicitly
enabled for the queue and each related messaged must contain the same session ID.
However, there are some important caveats: Service Bus queues size is limited to 80 GB, which is much smaller
than what's available from store queues. Additionally, Service Bus queues incur a base cost and charge per
operation.
Figure 4-14 outlines the high-level architecture of a Service Bus queue.
Figure 4-14 . Service Bus queue

In the previous figure, note the point-to-point relationship. Two instances of the same provider are enqueuing
messages into a single Service Bus queue. Each message is consumed by only one of three consumer instances on
the right. Next, we discuss how to implement messaging where different consumers may all be interested the same
message.
Events
Message queuing is an effective way to implement communication where a producer can asynchronously send a
consumer a message. However, what happens when many different consumers are interested in the same
message? A dedicated message queue for each consumer wouldn't scale well and would become difficult to
manage.
To address this scenario, we move to the third type of message interaction, the event. One microservice announces
that an action had occurred. Other microservices, if interested, react to the action, or event.
Eventing is a two-step process. For a given state change, a microservice publishes an event to a message broker,
making it available to any other interested microservice. The interested microservice is notified by subscribing to
the event in the message broker. You use the Publish/Subscribe pattern to implement event-based communication.
Figure 4-15 shows a shopping basket microservice publishing an event with two other microservices subscribing
to it.
Figure 4-15 . Event-Driven messaging
Note the event bus component that sits in the middle of the communication channel. It's a custom class that
encapsulates the message broker and decouples it from the underlying application. The ordering and inventory
microservices independently operate the event with no knowledge of each other, nor the shopping basket
microservice. When the registered event is published to the event bus, they act upon it.
With eventing, we move from queuing technology to topics. A topic is similar to a queue, but supports a one-to-
many messaging pattern. One microservice publishes a message. Multiple subscribing microservices can choose
to receive and act upon that message. Figure 4-16 shows a topic architecture.
Figure 4-16 . Topic architecture

In the previous figure, publishers send messages to the topic. At the end, subscribers receive messages from
subscriptions. In the middle, the topic forwards messages to subscriptions based on a set of rules, shown in dark
blue boxes. Rules act as a filter that forward specific messages to a subscription. Here, a "CreateOrder" event would
be sent to Subscription #1 and Subscription #3, but not to Subscription #2. An "OrderCompleted" event would be
sent to Subscription #2 and Subscription #3.
The Azure cloud supports two different topic services: Azure Service Bus Topics and Azure EventGrid.
Azure Service Bus Topics
Sitting on top of the same robust brokered message model of Azure Service Bus queues are Azure Service Bus
Topics. A topic can receive messages from multiple independent publishers and send messages to up to 2,000
subscribers. Subscriptions can be dynamically added or removed at runtime without stopping the system or
recreating the topic.
Many advanced features from Azure Service Bus queues are also available for topics, including Duplicate Detection
and Transaction support. By default, Service Bus topics are handled by a single message broker and stored in a
single message store. But, Service Bus Partitioning scales a topic by spreading it across many message brokers and
message stores.
Scheduled Message Delivery tags a message with a specific time for processing. The message won't appear in the
topic before that time. Message Deferral enables you to defer a retrieval of a message to a later time. Both are
commonly used in workflow processing scenarios where operations are processed in a particular order. You can
postpone processing of received messages until prior work has been completed.
Service Bus topics are a robust and proven technology for enabling publish/subscribe communication in your
cloud-native systems.
Azure Event Grid
While Azure Service Bus is a battle-tested messaging broker with a full set of enterprise features, Azure Event Grid
is the new kid on the block.
At first glance, Event Grid may look like just another topic-based messaging system. However, it's different in many
ways. Focused on event-driven workloads, it enables real-time event processing, deep Azure integration, and an
open-platform - all on serverless infrastructure. It's designed for contemporary cloud-native and serverless
applications
As a centralized eventing backplane, or pipe, Event Grid reacts to events inside Azure resources and from your own
services.
Event notifications are published to an Event Grid Topic, which, in turn, routes each event to a subscription.
Subscribers map to subscriptions and consume the events. Like Service Bus, Event Grid supports a filtered
subscriber model where a subscription sets rule for the events it wishes to receive. Event Grid provides fast
throughput with a guarantee of 10 million events per second enabling near real-time delivery - far more than what
Azure Service Bus can generate.
A sweet spot for Event Grid is its deep integration into the fabric of Azure infrastructure. An Azure resource, such as
Cosmos DB, can publish built-in events directly to other interested Azure resources - without the need for custom
code. Event Grid can publish events from an Azure Subscription, Resource Group, or Service, giving developers
fine-grained control over the lifecycle of cloud resources. However, Event Grid isn't limited to Azure. It's an open
platform that can consume custom HTTP events published from applications or third-party services and route
events to external subscribers.
When publishing and subscribing to native events from Azure resources, no coding is required. With simple
configuration, you can integrate events from one Azure resource to another leveraging built-in plumbing for Topics
and Subscriptions. Figure 4-17 shows the anatomy of Event Grid.
Figure 4-17 . Event Grid anatomy
A major difference between EventGrid and Service Bus is the underlying message exchange pattern.
Service Bus implements an older style pull model in which the downstream subscriber actively polls the topic
subscription for new messages. On the upside, this approach gives the subscriber full control of the pace at which
it processes messages. It controls when and how many messages to process at any given time. Unread messages
remain in the subscription until processed. A significant shortcoming is the latency between the time the event is
generated and the polling operation that pulls that message to the subscriber for processing. Also, the overhead of
constant polling for the next event consumes resources and money.
EventGrid, however, is different. It implements a push model in which events are sent to the EventHandlers as
received, giving near real-time event delivery. It also reduces cost as the service is triggered only when it's needed
to consume an event – not continually as with polling. That said, an event handler must handle the incoming load
and provide throttling mechanisms to protect itself from becoming overwhelmed. Many Azure services that
consume these events, such as Azure Functions and Logic Apps provide automatic autoscaling capabilities to
handle increased loads.
Event Grid is a fully managed serverless cloud service. It dynamically scales based on your traffic and charges you
only for your actual usage, not pre-purchased capacity. The first 100,000 operations per month are free –
operations being defined as event ingress (incoming event notifications), subscription delivery attempts,
management calls, and filtering by subject. With 99.99% availability, EventGrid guarantees the delivery of an event
within a 24-hour period, with built-in retry functionality for unsuccessful delivery. Undelivered messages can be
moved to a "dead-letter" queue for resolution. Unlike Azure Service Bus, Event Grid is tuned for fast performance
and doesn't support features like ordered messaging, transactions, and sessions.
Streaming messages in the Azure cloud
Azure Service Bus and Event Grid provide great support for applications that expose single, discrete events like a
new document has been inserted into a Cosmos DB. But, what if your cloud-native system needs to process a
stream of related events? Event streams are more complex. They're typically time-ordered, interrelated, and must
be processed as a group.
Azure Event Hub is a data streaming platform and event ingestion service that collects, transforms, and stores
events. It's fine-tuned to capture streaming data, such as continuous event notifications emitted from a telemetry
context. The service is highly scalable and can store and process millions of events per second. Shown in Figure 4-
18, it's often a front door for an event pipeline, decoupling ingest stream from event consumption.
Figure 4-18 . Azure Event Hub

Event Hub supports low latency and configurable time retention. Unlike queues and topics, Event Hubs keep event
data after it's been read by a consumer. This feature enables other data analytic services, both internal and external,
to replay the data for further analysis. Events stored in event hub are only deleted upon expiration of the retention
period, which is one day by default, but configurable.
Event Hub supports common event publishing protocols including HTTPS and AMQP. It also supports Kafka 1.0.
Existing Kafka applications can communicate with Event Hub using the Kafka protocol providing an alternative to
managing large Kafka clusters. Many open-source cloud-native systems embrace Kafka.
Event Hubs implements message streaming through a partitioned consumer model in which each consumer only
reads a specific subset, or partition, of the message stream. This pattern enables tremendous horizontal scale for
event processing and provides other stream-focused features that are unavailable in queues and topics. A partition
is an ordered sequence of events that is held in an event hub. As newer events arrive, they're added to the end of
this sequence. Figure 4-19 shows partitioning in an Event Hub.
Figure 4-19 . Event Hub partitioning

Instead of reading from the same resource, each consumer group reads across a subset, or partition, of the
message stream.
For cloud-native applications that must stream large numbers of events, Azure Event Hub can be a robust and
affordable solution.
PR EVIO U S NE XT
gRPC
So far in this book, we've focused on REST-based communication. We've seen that REST is a flexible architectural
style that defines CRUD-based operations against entity resources. Clients interact with resources across HTTP with
a request/response communication model. While REST is widely implemented, a newer communication
technology, gRPC, has gained tremendous momentum across the cloud-native community.
What is gRPC?
gRPC is a modern, high-performance framework that evolves the age-old remote procedure call (RPC) protocol. At
the application level, gRPC streamlines messaging between clients and back-end services. Originating from
Google, gRPC is open source and part of the Cloud Native Computing Foundation (CNCF) ecosystem of cloud-
native offerings. CNCF considers gRPC an incubating project. Incubating means end users are using the technology
in production applications, and the project has a healthy number of contributors.
A typical gRPC client app will expose a local, in-process function that implements a business operation. Under the
covers, that local function invokes another function on a remote machine. What appears to be a local call
essentially becomes a transparent out-of-process call to a remote service. The RPC plumbing abstracts the point-
to-point networking communication, serialization, and execution between computers.
In cloud-native applications, developers often work across programming languages, frameworks, and technologies.
This interoperability complicates message contracts and the plumbing required for cross-platform communication.
gRPC provides a "uniform horizontal layer" that abstracts these concerns. Developers code in their native platform
focused on business functionality, while gRPC handles communication plumbing.
gRPC offers comprehensive support across most popular development stacks, including Java, JavaScript, C#, Go,
Swift, and NodeJS.
gRPC Benefits
gRPC uses HTTP/2 for its transport protocol. While compatible with HTTP 1.1, HTTP/2 features many advanced
capabilities:
A binary protocol for data transport - unlike HTTP 1.1, which sends data as clear text.
Multiplexing support for sending multiple parallel requests over the same connection - HTTP 1.1 limits
processing to one request/response message at a time.
Bidirectional full-duplex communication for sending both client requests and server responses simultaneously.
Built-in streaming enabling requests and responses to asynchronously stream large data sets.
gRPC is lightweight and highly performant. It can be up to 8x faster than JSON serialization with messages 60-80%
smaller. In Microsoft Windows Communication Foundation (WCF) parlance, gRPC performance exceeds the speed
and efficiency of the highly optimized NetTCP bindings. Unlike NetTCP, which favors the Microsoft stack, gRPC is
cross-platform.
Protocol Buffers
gRPC embraces an open-source technology called Protocol Buffers. They provide a highly efficient and platform-
neutral serialization format for serializing structured messages that services send to each other. Using a cross-
platform Interface Definition Language (IDL), developers define a service contract for each microservice. The
contract, implemented as a text-based .proto file, describes the methods, inputs, and outputs for each service. The
same contract file can be used for gRPC clients and services built on different development platforms.
Using the proto file, the Protobuf compiler, protoc , generates both client and service code for your target
platform. The code includes the following components:
Strongly typed objects, shared by the client and service, that represent the service operations and data elements
for a message.
A strongly typed base class with the required network plumbing that the remote gRPC service can inherit and
extend.
A client stub that contains the required plumbing to invoke the remote gRPC service.
At runtime, each message is serialized as a standard Protobuf representation and exchanged between the client
and remote service. Unlike JSON or XML, Protobuf messages are serialized as compiled binary bytes.
The book, gRPC for WCF Developers, available from the Microsoft Architecture site, provides in-depth coverage of
gRPC and Protocol Buffers.
gRPC support in .NET

gRPC is integrated into .NET Core 3.0 SDK and later. The following tools support it:
Visual Studio 2019, version 16.3 or later, with the web development workload installed.
Visual Studio Code
the dotnet CLI
The SDK includes tooling for endpoint routing, built-in IoC, and logging. The open-source Kestrel web server
supports HTTP/2 connections. Figure 4-20 shows a Visual Studio 2019 template that scaffolds a skeleton project
for a gRPC service. Note how .NET Core fully supports Windows, Linux, and macOS.
Figure 4-20 . gRPC support in Visual Studio 2019

Figure 4-21 shows the skeleton gRPC service generated from the built-in scaffolding included in Visual Studio
2019.
Figure 4-21 . gRPC project in Visual Studio 2019
In the previous figure, note the proto description file and service code. As you'll see shortly, Visual Studio generates
additional configuration in both the Startup class and underlying project file.
gRPC usage
Favor gRPC for the following scenarios:
Synchronous backend microservice-to-microservice communication where an immediate response is required
to continue processing.
Polyglot environments that need to support mixed programming platforms.
Low latency and high throughput communication where performance is critical.
Point-to-point real-time communication - gRPC can push messages in real time without polling and has
excellent support for bi-directional streaming.
Network constrained environments – binary gRPC messages are always smaller than an equivalent text-based
JSON message.
At the time, of this writing, gRPC is primarily used with backend services. Most modern browsers can't provide the
level of HTTP/2 control required to support a front-end gRPC client. That said, there's an early initiative that enables
gRPC communication from browser-based apps built with JavaScript or Blazor WebAssembly technologies. The
gRPC-Web for .NET enables an ASP.NET Core gRPC app to support gRPC features in browser apps:
Strongly typed, code-generated clients
Compact Protobuf messages
Server streaming
gRPC implementation
The microservice reference architecture, eShop on Containers, from Microsoft, shows how to implement gRPC
services in .NET Core applications. Figure 4-22 presents the back-end architecture.
Figure 4-22 . Backend architecture for eShop on Containers
In the previous figure, note how eShop embraces the Backend for Frontends pattern (BFF) by exposing multiple API
gateways. We discussed the BFF pattern earlier in this chapter. Pay close attention to the Aggregator microservice
(in gray) that sits between the Web-Shopping API Gateway and backend Shopping microservices. The Aggregator
receives a single request from a client, dispatches it to various microservices, aggregates the results, and sends
them back to the requesting client. Such operations typically require synchronous communication as to produce an
immediate response. In eShop, backend calls from the Aggregator are performed using gRPC as shown in Figure
4-23.
Figure 4-23 . gRPC in eShop on Containers

gRPC communication requires both client and server components. In the previous figure, note how the Shopping
Aggregator implements a gRPC client. The client makes synchronous gRPC calls (in red) to backend microservices,
each of which implement a gRPC server. Both the client and server take advantage of the built-in gRPC plumbing
from the .NET Core SDK. Client-side stubs provide the plumbing to invoke remote gRPC calls. Server-side
components provide gRPC plumbing that custom service classes can inherit and consume.
Microservices that expose both a RESTful API and gRPC communication require multiple endpoints to manage
traffic. You would open an endpoint that listens for HTTP traffic for the RESTful calls and another for gRPC calls. The
gRPC endpoint must be configured for the HTTP/2 protocol that is required for gRPC communication.
While we strive to decouple microservices with asynchronous communication patterns, some operations require
direct calls. gRPC should be the primary choice for direct synchronous communication between microservices. Its
high-performance communication protocol, based on HTTP/2 and protocol buffers, make it a perfect choice.
Looking ahead
Looking ahead, gRPC will continue to gain traction for cloud-native systems. The performance benefits and ease of
development are compelling. However, REST will likely be around for a long time. It excels for publicly exposed APIs
and for backward compatibility reasons.
PR EVIO U S NE XT
Service Mesh communication infrastructure
Throughout this chapter, we've explored the challenges of microservice communication. We said that development
teams need to be sensitive to how back-end services communicate with each other. Ideally, the less inter-service
communication, the better. However, avoidance isn't always possible as back-end services often rely on one another
to complete operations.
We explored different approaches for implementing synchronous HTTP communication and asynchronous
messaging. In each of the cases, the developer is burdened with implementing communication code.
Communication code is complex and time intensive. Incorrect decisions can lead to significant performance issues.
A more modern approach to microservice communication centers around a new and rapidly evolving technology
entitled Service Mesh. A service mesh is a configurable infrastructure layer with built-in capabilities to handle
service-to-service communication, resiliency, and many cross-cutting concerns. It moves the responsibility for
these concerns out of the microservices and into service mesh layer. Communication is abstracted away from your
microservices.
A key component of a service mesh is a proxy. In a cloud-native application, an instance of a proxy is typically
colocated with each microservice. While they execute in separate processes, the two are closely linked and share
the same lifecycle. This pattern, known as the Sidecar pattern, and is shown in Figure 4-24.
Figure 4-24 . Service mesh with a side car

Note in the previous figure how messages are intercepted by a proxy that runs alongside each microservice. Each
proxy can be configured with traffic rules specific to the microservice. It understands messages and can route them
across your services and the outside world.
Along with managing service-to-service communication, the Service Mesh provides support for service discovery
and load balancing.
Once configured, a service mesh is highly functional. The mesh retrieves a corresponding pool of instances from a
service discovery endpoint. It sends a request to a specific service instance, recording the latency and response
type of the result. It chooses the instance most likely to return a fast response based on different factors, including
the observed latency for recent requests.
A service mesh manages traffic, communication, and networking concerns at the application level. It understands
messages and requests. A service mesh typically integrates with a container orchestrator. Kubernetes supports an
extensible architecture in which a service mesh can be added.
In chapter 6, we deep-dive into Service Mesh technologies including a discussion on its architecture and available
open-source implementations.
Summary
In this chapter, we discussed cloud-native communication patterns. We started by examining how front-end clients
communicate with back-end microservices. Along the way, we talked about API Gateway platforms and real-time
communication. We then looked at how microservices communicate with other back-end services. We looked at
both synchronous HTTP communication and asynchronous messaging across services. We covered gRPC, an
upcoming technology in the cloud-native world. Finally, we introduced a new and rapidly evolving technology
entitled Service Mesh that can streamline microservice communication.
Special emphasis was on managed Azure services that can help implement communication in cloud-native
systems:
Azure Application Gateway
Azure SignalR Service
Azure Storage Queues
Azure Service Bus
Azure Event Grid
Azure Event Hub
We next move to distributed data in cloud-native systems and the benefits and challenges that it presents.
References
.NET Microservices: Architecture for Containerized .NET applications
Designing Interservice Communication for Microservices
Azure SignalR Service, a fully managed service to add real-time functionality
Azure API Gateway Ingress Controller
About Ingress in Azure Kubernetes Service (AKS)
gRPC Documentation
gRPC for WCF Developers
Comparing gRPC Services with HTTP APIs
Building gRPC Services with .NET video
PR EVIO U S NE XT
Distributed data
As we've seen throughout this book, a cloud-native approach changes the way you design, deploy, and manage
applications. It also changes the way you manage and store data.
Figure 5-1 contrasts the differences.
Figure 5-1 . Data management in cloud-native applications

Experienced developers will easily recognize the architecture on the left-side of figure 5-1. In this monolithic
application, business service components collocate together in a shared services tier, sharing data from a single
relational database.
In many ways, a single database keeps data management simple. Querying data across multiple tables is
straightforward. Changes to data update together or they all rollback. ACID transactions guarantee strong and
immediate consistency.
Designing for cloud-native, we take a different approach. On the right-side of Figure 5-1, note how business
functionality segregates into small, independent microservices. Each microservice encapsulates a specific business
capability and its own data. The monolithic database decomposes into a distributed data model with many smaller
databases, each aligning with a microservice. When the smoke clears, we emerge with a design that exposes a
database per microservice.
Database-per-microservice, why?
This database per microservice provides many benefits, especially for systems that must evolve rapidly and
support massive scale. With this model...
Domain data is encapsulated within the service
Data schema can evolve without directly impacting other services
Each data store can independently scale
A data store failure in one service won't directly impact other services
Segregating data also enables each microservice to implement the data store type that is best optimized for its
workload, storage needs, and read/write patterns. Choices include relational, document, key-value, and even
graph-based data stores.
Figure 5-2 presents the principle of polyglot persistence in a cloud-native system.
Figure 5-2 . Polyglot data persistence

Note in the previous figure how each microservice supports a different type of data store.
The product catalog microservice consumes a relational database to accommodate the rich relational structure
of its underlying data.
The shopping cart microservice consumes a distributed cache that supports its simple, key-value data store.
The ordering microservice consumes both a NoSql document database for write operations along with a highly
denormalized key/value store to accommodate high-volumes of read operations.
While relational databases remain relevant for microservices with complex data, NoSQL databases have gained
considerable popularity. They provide massive scale and high availability. Their schemaless nature allows
developers to move away from an architecture of typed data classes and ORMs that make change expensive and
time-consuming. We cover NoSQL databases later in this chapter.
While encapsulating data into separate microservices can increase agility, performance, and scalability, it also
presents many challenges. In the next section, we discuss these challenges along with patterns and practices to
help overcome them.
Cross-service queries
While microservices are independent and focus on specific functional capabilities, like inventory, shipping, or
ordering, they frequently require integration with other microservices. Often the integration involves one
microservice querying another for data. Figure 5-3 shows the scenario.
Figure 5-3 . Querying across microservices
In the preceding figure, we see a shopping basket microservice that adds an item to a user's shopping basket.
While the data store for this microservice contains basket and line item data, it doesn't maintain product or pricing
data. Instead, those data items are owned by the catalog and pricing microservices. This presents a problem. How
can the shopping basket microservice add a product to the user's shopping basket when it doesn't have product
nor pricing data in its database?
One option discussed in Chapter 4 is a direct HTTP call from the shopping basket to the catalog and pricing
microservices. However, in chapter 4, we said synchronous HTTP calls couple microservices together, reducing
their autonomy and diminishing their architectural benefits.
We could also implement a request-reply pattern with separate inbound and outbound queues for each service.
However, this pattern is complicated and requires plumbing to correlate request and response messages. While it
does decouple the backend microservice calls, the calling service must still synchronously wait for the call to
complete. Network congestion, transient faults, or an overloaded microservice and can result in long-running and
even failed operations.
Instead, a widely accepted pattern for removing cross-service dependencies is the Materialized View Pattern,
shown in Figure 5-4.
Figure 5-4 . Materialized View Pattern

With this pattern, you place a local data table (known as a read model) in the shopping basket service. This table
contains a denormalized copy of the data needed from the product and pricing microservices. Copying the data
directly into the shopping basket microservice eliminates the need for expensive cross-service calls. With the data
local to the service, you improve the service's response time and reliability. Additionally, having its own copy of the
data makes the shopping basket service more resilient. If the catalog service should become unavailable, it
wouldn't directly impact the shopping basket service. The shopping basket can continue operating with the data
from its own store.
The catch with this approach is that you now have duplicate data in your system. However, strategically duplicating
data in cloud-native systems is an established practice and not considered an anti-pattern, or bad practice. Keep in
mind that one and only one service can own a data set and have authority over it. You'll need to synchronize the
read models when the system of record is updated. Synchronization is typically implemented via asynchronous
messaging with a publish/subscribe pattern, as shown in Figure 5.4.
Distributed transactions
While querying data across microservices is difficult, implementing a transaction across several microservices is
even more complex. The inherent challenge of maintaining data consistency across independent data sources in
different microservices can't be understated. The lack of distributed transactions in cloud-native applications
means that you must manage distributed transactions programmatically. You move from a world of immediate
consistency to that of eventual consistency.
Figure 5-5 shows the problem.
Figure 5-5 . Implementing a transaction across microservices

In the preceding figure, five independent microservices participate in a distributed transaction that creates an
order. Each microservice maintains its own data store and implements a local transaction for its store. To create the
order, the local transaction for each individual microservice must succeed, or all must abort and roll back the
operation. While built-in transactional support is available inside each of the microservices, there's no support for
a distributed transaction that would span across all five services to keep data consistent.
Instead, you must construct this distributed transaction programmatically.
A popular pattern for adding distributed transactional support is the Saga pattern. It's implemented by grouping
local transactions together programmatically and sequentially invoking each one. If any of the local transactions
fail, the Saga aborts the operation and invokes a set of compensating transactions. The compensating transactions
undo the changes made by the preceding local transactions and restore data consistency. Figure 5-6 shows a failed
transaction with the Saga pattern.
Figure 5-6 . Rolling back a transaction
In the previous figure, the Update Inventory operation has failed in the Inventory microservice. The Saga invokes a
set of compensating transactions (in red) to adjust the inventory counts, cancel the payment and the order, and
return the data for each microservice back to a consistent state.
Saga patterns are typically choreographed as a series of related events, or orchestrated as a set of related
commands. In Chapter 4, we discussed the service aggregator pattern that would be the foundation for an
orchestrated saga implementation. We also discussed eventing along with Azure Service Bus and Azure Event Grid
topics that would be a foundation for a choreographed saga implementation.
High volume data

Large cloud-native applications often support high-volume data requirements. In these scenarios, traditional data
storage techniques can cause bottlenecks. For complex systems that deploy on a large scale, both Command and
Query Responsibility Segregation (CQRS) and Event Sourcing may improve application performance.
CQRS
CQRS, is an architectural pattern that can help maximize performance, scalability, and security. The pattern
separates operations that read data from those operations that write data.
For normal scenarios, the same entity model and data repository object are used for both read and write
operations.
However, a high volume data scenario can benefit from separate models and data tables for reads and writes. To
improve performance, the read operation could query against a highly denormalized representation of the data to
avoid expensive repetitive table joins and table locks. The write operation, known as a command, would update
against a fully normalized representation of the data that would guarantee consistency. You then need to
implement a mechanism to keep both representations in sync. Typically, whenever the write table is modified, it
publishes an event that replicates the modification to the read table.
Figure 5-7 shows an implementation of the CQRS pattern.
Figure 5-7 . CQRS implementation
In the previous figure, separate command and query models are implemented. Each data write operation is saved
to the write store and then propagated to the read store. Pay close attention to how the data propagation process
operates on the principle of eventual consistency. The read model eventually synchronizes with the write model,
but there may be some lag in the process. We discuss eventual consistency in the next section.
This separation enables reads and writes to scale independently. Read operations use a schema optimized for
queries, while the writes use a schema optimized for updates. Read queries go against denormalized data, while
complex business logic can be applied to the write model. As well, you might impose tighter security on write
operations than those exposing reads.
Implementing CQRS can improve application performance for cloud-native services. However, it does result in a
more complex design. Apply this principle carefully and strategically to those sections of your cloud-native
application that will benefit from it. For more on CQRS, see the Microsoft book .NET Microservices: Architecture for
Containerized .NET Applications.
Event sourcing
Another approach to optimizing high volume data scenarios involves Event Sourcing.
A system typically stores the current state of a data entity. If a user changes their phone number, for example, the
customer record is updated with the new number. We always know the current state of a data entity, but each
update overwrites the previous state.
In most cases, this model works fine. In high volume systems, however, overhead from transactional locking and
frequent update operations can impact database performance, responsiveness, and limit scalability.
Event Sourcing takes a different approach to capturing data. Each operation that affects data is persisted to an
event store. Instead of updating the state of a data record, we append each change to a sequential list of past
events - similar to an accountant's ledger. The Event Store becomes the system of record for the data. It's used to
propagate various materialized views within the bounded context of a microservice. Figure 5.8 shows the pattern.
Figure 5-8 . Event Sourcing
In the previous figure, note how each entry (in blue) for a user's shopping cart is appended to an underlying event
store. In the adjoining materialized view, the system projects the current state by replaying all the events
associated with each shopping cart. This view, or read model, is then exposed back to the UI. Events can also be
integrated with external systems and applications or queried to determine the current state of an entity. With this
approach, you maintain history. You know not only the current state of an entity, but also how you reached this
state.
Mechanically speaking, event sourcing simplifies the write model. There are no updates or deletes. Appending each
data entry as an immutable event minimizes contention, locking, and concurrency conflicts associated with
relational databases. Building read models with the materialized view pattern enables you to decouple the view
from the write model and choose the best data store to optimize the needs of your application UI.
For this pattern, consider a data store that directly supports event sourcing. Azure Cosmos DB, MongoDB,
Cassandra, CouchDB, and RavenDB are good candidates.
As with all patterns and technologies, implement strategically and when needed. While event sourcing can provide
increased performance and scalability, it comes at the expense of complexity and a learning curve.
PR EVIO U S NE XT
Relational vs. NoSQL data
Relational and NoSQL are two types of database systems commonly implemented in cloud-native apps. They're
built differently, store data differently, and accessed differently. In this section, we'll look at both. Later in this
chapter, we'll look at an emerging database technology called NewSQL.
Relational databases have been a prevalent technology for decades. They're mature, proven, and widely
implemented. Competing database products, tooling, and expertise abound. Relational databases provide a store of
related data tables. These tables have a fixed schema, use SQL (Structured Query Language) to manage data, and
support ACID guarantees.
No-SQL databases refer to high-performance, non-relational data stores. They excel in their ease-of-use, scalability,
resilience, and availability characteristics. Instead of joining tables of normalized data, NoSQL stores unstructured
or semi-structured data, often in key-value pairs or JSON documents. No-SQL databases typically don't provide
ACID guarantees beyond the scope of a single database partition. High volume services that require sub second
response time favor NoSQL datastores.
The impact of NoSQL technologies for distributed cloud-native systems can't be overstated. The proliferation of
new data technologies in this space has disrupted solutions that once exclusively relied on relational databases.
NoSQL databases include several different models for accessing and managing data, each suited to specific use
cases. Figure 5-9 presents four common models.
Figure 5-9 : Data models for NoSQL databases
M O DEL C H A RA C T ERIST IC S
Document Store Data and metadata are stored hierarchically in JSON-based

documents inside the database.
Key Value Store The simplest of the NoSQL databases, data is represented as a
collection of key-value pairs.
Wide-Column Store Related data is stored as a set of nested-key/value pairs within

a single column.
Graph Store Data is stored in a graph structure as node, edge, and data
properties.
The CAP theorem

As a way to understand the differences between these types of databases, consider the CAP theorem, a set of
principles applied to distributed systems that store state. Figure 5-10 shows the three properties of the CAP
theorem.
Figure 5-10 . The CAP theorem

The theorem states that distributed data systems will offer a trade-off between consistency, availability, and
partition tolerance. And, that any database can only guarantee two of the three properties:
Consistency. Every node in the cluster responds with the most recent data, even if the system must block the
request until all replicas update. If you query a "consistent system" for an item that is currently updating,
you'll wait for that response until all replicas successfully update. However, you'll receive the most current
data.
Availability. Every node returns an immediate response, even if that response isn't the most recent data. If
you query an "available system" for an item that is updating, you'll get the best possible answer the service
can provide at that moment.
Partition Tolerance. Guarantees the system continues to operate even if a replicated data node fails or loses
connectivity with other replicated data nodes.
Relational databases typically provide consistency and availability, but not partition tolerance. They're typically
provisioned to a single server and scale vertically by adding more resources to the machine.
Many relational database systems support built-in replication features where copies of the primary database can
be made to other secondary server instances. Write operations are made to the primary instance and replicated to
each of the secondaries. Upon a failure, the primary instance can fail over to a secondary to provide high
availability. Secondaries can also be used to distribute read operations. While writes operations always go against
the primary replica, read operations can be routed to any of the secondaries to reduce system load.
Data can also be horizontally partitioned across multiple nodes, such as with sharding. But, sharding dramatically
increases operational overhead by spitting data across many pieces that cannot easily communicate. It can be
costly and time consuming to manage. It can end up impacting performance, table joins, and referential integrity.
If data replicas were to lose network connectivity in a "highly consistent" relational database cluster, you wouldn't
be able to write to the database. The system would reject the write operation as it can't replicate that change to the
other data replica. Every data replica has to update before the transaction can complete.
NoSQL databases typically support high availability and partition tolerance. They scale out horizontally, often
across commodity servers. This approach provides tremendous availability, both within and across geographical
regions at a reduced cost. You partition and replicate data across these machines, or nodes, providing redundancy
and fault tolerance. The downside is consistency. A change to data on one NoSQL node can take some time to
propagate to other nodes. Typically, a NoSQL database node will provide an immediate response to a query - even
if the data that is presented is stale and hasn't updated yet.
If data replicas were to lose connectivity in a "highly available" NoSQL database cluster, you could still complete a
write operation to the database. The database cluster would allow the write operation and update each data replica
as it becomes available.
This kind of result is known as eventual consistency, a characteristic of distributed data systems where ACID
transactions aren't supported. It's a brief delay between the update of a data item and time that it takes to
propagate that update to each of the replica nodes. Under normal conditions, the lag is typically short, but can
increase when problems arise. For example, what would happen if you were to update a product item in a NoSQL
database in the United States and query that same data item from a replica node in Europe? You would receive the
earlier product information, until the cluster updates the European node with the product change. By immediately
returning a query result and not waiting for all replica nodes to update, you gain enormous scale and volume, but
with the possibility of presenting older data.
High availability and massive scalability are often more critical to the business than strong consistency. Developers
can implement techniques and patterns such as Sagas, CQRS, and asynchronous messaging to embrace eventual
consistency.
Nowadays, care must be taken when conidering the CAP theorem constraints. A new type of database, called
NewSQL, has emerged which extends the relational database engine to support both horizontal scalability and
the scalable performance of NoSQL systems.
Considerations for relational vs. NoSQL systems

Based upon specific data requirements, a cloud-native-based microservice can implement a relational, NoSQL
datastore or both.
C O N SIDER A N O SQ L DATA STO RE W H EN : C O N SIDER A REL AT IO N A L DATA B A SE W H EN :
You have high volume workloads that require large scale Your workload volume is consistent and requires medium to
large scale
Your workloads don't require ACID guarantees ACID guarantees are required
Your data is dynamic and frequently changes Your data is predictable and highly structured
Data can be expressed without relationships Data is best expressed relationally
You need fast writes and write safety isn't critical Write safety is a requirement
Data retrieval is simple and tends to be flat You work with complex queries and reports
Your data requires a wide geographic distribution Your users are more centralized
Your application will be deployed to commodity hardware, Your application will be deployed to large, high-end hardware
such as with public clouds
In the next sections, we'll explore the options available in the Azure cloud for storing and managing your cloud-
native data.
Database as a Service
To start, you could provision an Azure virtual machine and install your database of choice for each service. While
you'd have full control over the environment, you'd forgo many built-in features of the cloud platform. You'd also
be responsible for managing the virtual machine and database for each service. This approach could quickly
become time-consuming and expensive.
Instead, cloud-native applications favor data services exposed as a Database as a Service (DBaaS). Fully managed
by a cloud vendor, these services provide built-in security, scalability, and monitoring. Instead of owning the
service, you simply consume it as a backing service. The provider operates the resource at scale and bears the
responsibility for performance and maintenance.
They can be configured across cloud availability zones and regions to achieve high availability. They all support
just-in-time capacity and a pay-as-you-go model. Azure features different kinds of managed data service options,
each with specific benefits.
We'll first look at relational DBaaS services available in Azure. You'll see that Microsoft's flagship SQL Server
database is available along with several open-source options. Then, we'll talk about the NoSQL data services in
Azure.
Azure relational databases

For cloud-native microservices that require relational data, Azure offers four managed relational databases as a
service (DBaaS) offerings, shown in Figure 5-11.
Figure 5-11 . Managed relational databases available in Azure

In the previous figure, note how each sits upon a common DBaaS infrastructure which features key capabilities at
no additional cost.
These features are especially important to organizations who provision large numbers of databases, but have
limited resources to administer them. You can provision an Azure database in minutes by selecting the amount of
processing cores, memory, and underlying storage. You can scale the database on-the-fly and dynamically adjust
resources with little to no downtime.
Azure SQL Database

Development teams with expertise in Microsoft SQL Server should consider Azure SQL Database. It's a fully
managed relational database-as-a-service (DBaaS) based on the Microsoft SQL Server Database Engine. The
service shares many features found in the on-premises version of SQL Server and runs the latest stable version of
the SQL Server Database Engine.
For use with a cloud-native microservice, Azure SQL Database is available with three deployment options:
A Single Database represents a fully managed SQL Database running on an Azure SQL Database server in
the Azure cloud. The database is considered contained as it has no configuration dependencies on the
underlying database server.
A Managed Instance is a fully managed instance of the Microsoft SQL Server Database Engine that provides
near-100% compatibility with an on-premises SQL Server. This option supports larger databases, up to 35
TB and is placed in an Azure Virtual Network for better isolation.
Azure SQL Database serverless is a compute tier for a single database that automatically scales based on
workload demand. It bills only for the amount of compute used per second. The service is well suited for
workloads with intermittent, unpredictable usage patterns, interspersed with periods of inactivity. The
serverless compute tier also automatically pauses databases during inactive periods so that only storage
charges are billed. It automatically resumes when activity returns.
Beyond the traditional Microsoft SQL Server stack, Azure also features managed versions of three popular open-
source databases.
Open-source databases in Azure

Open-source relational databases have become a popular choice for cloud-native applications. Many enterprises
favor them over commercial database products, especially for cost savings. Many development teams enjoy their
flexibility, community-backed development, and ecosystem of tools and extensions. Open-source databases can be
deployed across multiple cloud providers, helping minimize the concern of "vendor lock-in."
Developers can easily self-host any open-source database on an Azure VM. While providing full control, this
approach puts you on the hook for the management, monitoring, and maintenance of the database and VM.
However, Microsoft continues its commitment to keeping Azure an “open platform” by offering several popular
open-source databases as fully managed DBaaS services.
Azure Database for MySQL
MySQL is an open-source relational database and a pillar for applications built on the LAMP software stack. Widely
chosen for read heavy workloads, it's used by many large organizations, including Facebook, Twitter, and YouTube.
The community edition is available for free, while the enterprise edition requires a license purchase. Originally
created in 1995, the product was purchased by Sun Microsystems in 2008. Oracle acquired Sun and MySQL in
2010.
Azure Database for MySQL is a managed relational database service based on the open-source MySQL Server
engine. It uses the MySQL Community edition. The Azure MySQL server is the administrative point for the service.
It's the same MySQL server engine used for on-premises deployments. The engine can create a single database per
server or multiple databases per server that share resources. You can continue to manage data using the same
open-source tools without having to learn new skills or manage virtual machines.
Azure Database for MariaDB
MariaDB Server is another popular open-source database server. It was created as a fork of MySQL when Oracle
purchased Sun Microsystems, who owned MySQL. The intent was to ensure that MariaDB remained open-source.
As MariaDB is a fork of MySQL, the data and table definitions are compatible, and the client protocols, structures,
and APIs, are close-knit.
MariaDB has a strong community and is used by many large enterprises. While Oracle continues to maintain,
enhance, and support MySQL, the MariaDB foundation manages MariaDB, allowing public contributions to the
product and documentation.
Azure Database for MariaDB is a fully managed relational database as a service in the Azure cloud. The service is
based on the MariaDB community edition server engine. It can handle mission-critical workloads with predictable
performance and dynamic scalability.
Azure Database for PostgreSQL
PostgreSQL is an open-source relational database with over 30 years of active development. PostgresSQL has a
strong reputation for reliability and data integrity. It’s feature rich, SQL compliant, and considered more performant
than MySQL - especially for workloads with complex queries and heavy writes. Many large enterprises including
Apple, Red Hat, and Fujitsu have built products using PostgreSQL.
Azure Database for PostgreSQL is a fully managed relational database service, based on the open-source Postgres
database engine. The service supports many development platforms, including C++, Java, Python, Node, C#, and
PHP. You can migrate PostgreSQL databases to it using the command-line interface tool or Azure Data Migration
Service.
Azure Database for PostgreSQL is available with two deployment options:
The Single Server deployment option is a central administrative point for multiple databases to which you
can deploy many databases. The pricing is structured per-server based upon cores and storage.
The Hyperscale (Citus) option is powered by Citus Data technology. It enables high performance by
horizontally scaling a single database across hundreds of nodes to deliver fast performance and scale. This
option allows the engine to fit more data in memory, parallelize queries across hundreds of nodes, and
index data faster.
NoSQL data in Azure

Cosmos DB is a fully managed, globally distributed NoSQL database service in the Azure cloud. It has been
adopted by many large companies across the world, including Coca-Cola, Skype, ExxonMobil, and Liberty Mutual.
If your services require fast response from anywhere in the world, high availability, or elastic scalability, Cosmos DB
is a great choice. Figure 5-12 shows Cosmos DB.
Figure 5-12 : Overview of Cosmos DB

The previous figure presents many of the built-in cloud-native capabilities available in Cosmos DB. In this section,
we’ll take a closer look at them.
Global support
Cloud-native applications often have a global audience and require global scale.
You can distribute Cosmos databases across regions or around the world, placing data close to your users,
improving response time, and reducing latency. You can add or remove a database from a region without pausing
or redeploying your services. In the background, Cosmos DB transparently replicates the data to each of the
configured regions.
Cosmos DB supports active/active clustering at the global level, enabling you to configure any of your database
regions to support both writes and reads.
The Multi-Master protocol is an important feature in Cosmos DB that enables the following functionality:
Unlimited elastic write and read scalability.
99.999% read and write availability all around the world.
Guaranteed reads and writes served in less than 10 milliseconds at the 99th percentile.
With the Cosmos DB Multi-Homing APIs, your microservice is automatically aware of the nearest Azure region and
sends requests to it. The nearest region is identified by Cosmos DB without any configuration changes. Should a
region become unavailable, the Multi-Homing feature will automatically route requests to the next nearest
available region.
Multi-model support
When replatforming monolithic applications to a cloud-native architecture, development teams sometimes have to
migrate open-source, NoSQL data stores. Cosmos DB can help you preserve your investment in these NoSQL
datastores with its multi-model data platform. The following table shows the supported NoSQL compatibility APIs.
P RO VIDER DESC RIP T IO N
SQL API Proprietary API that supports JSON documents and SQL-
based queries
Mongo DB API Supports Mongo DB APIs and JSON documents
Gremlin API Supports Gremlin API with graph-based nodes and edge data
representations
Cassandra API Supports Casandra API for wide-column data representations
Table API Supports Azure Table Storage with premium enhancements
etcd API Enables Cosmos DB as a backing store for Azure Kubernetes

Service clusters
Development teams can migrate existing Mongo, Gremlin, or Cassandra databases into Cosmos DB with minimal
changes to data or code. For new apps, development teams can choose among open-source options or the built-in
SQL API model.
Internally, Cosmos stores the data in a simple struct format made up of primitive data types. For each request,
the database engine translates the primitive data into the model representation you've selected.
In the previous table, note the Table API option. This API is an evolution of Azure Table Storage. Both share the same
underlying table model, but the Cosmos DB Table API adds premium enhancements not available in the Azure
Storage API. The following table contrasts the features.
A Z URE TA B L E STO RA GE A Z URE C O SM O S DB
Latency Fast Single-digit millisecond latency for reads

and writes anywhere in the world
Throughput Limit of 20,000 operations per table 10 Million operations per table
Global Distribution Single region with optional single Turnkey distributions to all regions with
secondary read region automatic failover
Indexing Available for partition and row key Automatic indexing of all properties
properties only
Pricing Based on storage Based on throughput
Microservices that consume Azure Table storage can easily migrate to the Cosmos DB Table API. No code changes
are required.
Tunable consistency
Earlier in the Relational vs. NoSQL section, we discussed the subject of data consistency. Data consistency refers to
the integrity of your data. Cloud-native services with distributed data rely on replication and must make a
fundamental tradeoff between read consistency, availability, and latency.
Most distributed databases allow developers to choose between two consistency models: strong consistency
and eventual consistency. Strong consistency is the gold standard of data programmability. It guarantees that a
query will always return the most current data - even if the system must incur latency waiting for an update to
replicate across all database copies. While a database configured for eventual consistency will return data
immediately, even if that data isn't the most current copy. The latter option enables higher availability, greater scale,
and increased performance.
Azure Cosmos DB offers five well-defined consistency models shown in Figure 5-13.
Figure 5-13 : Cosmos DB Consistency Levels

These options enable you to make precise choices and granular tradeoffs for consistency, availability, and the
performance for your data. The levels are presented in the following table.
C O N SIST EN C Y L EVEL DESC RIP T IO N
Eventual No ordering guarantee for reads. Replicas will eventually

converge.
Constant Prefix Reads are still eventual, but data is returned in the ordering in
which it is written.
Session Guarantees you can read any data written during the current
session. It is the default consistency level.
Bounded Staleness Reads trail writes by interval that you specify.

C O N SIST EN C Y L EVEL DESC RIP T IO N
Strong Reads are guaranteed to return most recent committed

version of an item. A client never sees an uncommitted or
partial read.
In the article Getting Behind the 9-Ball: Cosmos DB Consistency Levels Explained, Microsoft Program Manager
Jeremy Likness provides an excellent explanation of the five models.
Partitioning
Azure Cosmos DB embraces automatic partitioning to scale a database to meet the performance needs of your
cloud-native services.
You manage data in Cosmos DB data by creating databases, containers, and items.
Containers live in a Cosmos DB database and represent a schema-agnostic grouping of items. Items are the data
that you add to the container. They're represented as documents, rows, nodes, or edges. All items added to a
container are automatically indexed.
To partition the container, items are divided into distinct subsets called logical partitions. Logical partitions are
populated based on the value of a partition key that is associated with each item in a container. Figure 5-14 shows
two containers each with a logical partition based on a partition key value.
Figure 5-14 : Cosmos DB partitioning mechanics

Note in the previous figure how each item includes a partition key of either ‘city’ or ‘airport’. The key determines
the item’s logical partition. Items with a city code are assigned to the container on the left, and items with an
airport code, to the container on the right. Combining the partition key value with the ID value creates an
item's index, which uniquely identifies the item.
Internally, Cosmos DB automatically manages the placement of logical partitions on physical partitions to satisfy
the scalability and performance needs of the container. As application throughput and storage requirements
increase, Azure Cosmos DB redistributes logical partitions across a greater number of servers. Redistribution
operations are managed by Cosmos DB and invoked without interruption or downtime.
NewSQL databases
NewSQL is an emerging database technology that combines the distributed scalability of NoSQL with the
ACID guarantees of a relational database. NewSQL databases are important for business systems that must
process high-volumes of data, across distributed environments, with full transactional support and ACID
compliance. While a NoSQL database can provide massive scalability, it does not guarantee data consistency.
Intermittent problems from inconsistent data can place a burden on the development team. Developers must
construct safeguards into their microservice code to manage problems caused by inconsistent data.
The Cloud Native Computing Foundation (CNCF) features several NewSQL database projects.
P RO JEC T C H A RA C T ERIST IC S
Cockroach DB An ACID-compliant, relational database that scales globally.

Add a new node to a cluster and CockroachDB takes care of
balancing the data across instances and geographies. It
creates, manages, and distributes replicas to ensure reliability.
It’s open source and freely available.
TiDB An open-source database that supports Hybrid Transactional

and Analytical Processing (HTAP) workloads. It is MySQL-
compatible and features horizontal scalability, strong
consistency, and high availability. TiDB acts like a MySQL
server. You can continue to use existing MySQL client libraries,
without requiring extensive code changes to your application.
YugabyteDB An open source, high-performance, distributed SQL database.

It supports low query latency, resilience against failures, and
global data distribution. YugabyteDB is PostgressSQL-
compatible and handles scale-out RDBMS and internet-scale
OLTP workloads. The product also supports NoSQL and is
compatible with Cassandra.
Vitess Vitess is a database solution for deploying, scaling, and

managing large clusters of MySQL instances. It can run in a
public or private cloud architecture. Vitess combines and
extends many important MySQL features and features both
vertical and horizontal sharding support. Originated by
YouTube, Vitess has been serving all YouTube database traffic
since 2011.
The open-source projects in the previous figure are available from the Cloud Native Computing Foundation. Three
of the offerings are full database products, which include .NET Core support. The other, Vitess, is a database
clustering system that horizontally scales large clusters of MySQL instances.
A key design goal for NewSQL databases is to work natively in Kubernetes, taking advantage of the platform's
resiliency and scalability.
NewSQL databases are designed to thrive in ephemeral cloud environments where underlying virtual machines
can be restarted or rescheduled at a moment’s notice. The databases are designed to survive node failures without
data loss nor downtime. CockroachDB, for example, is able to survive a machine loss by maintaining three
consistent replicas of any data across the nodes in a cluster.
Kubernetes uses a Services construct to allow a client to address a group of identical NewSQL databases processes
from a single DNS entry. By decoupling the database instances from the address of the service with which it's
associated, we can scale without disrupting existing application instances. Sending a request to any service at a
given time will always yield the same result.
In this scenario, all database instances are equal. There are no primary or secondary relationships. Techniques like
consensus replication found in CockroachDB allow any database node to handle any request. If the node that
receives a load-balanced request has the data it needs locally, it responds immediately. If not, the node becomes a
gateway and forwards the request to the appropriate nodes to get the correct answer. From the client's perspective,
every database node is the same: They appear as a single logical database with the consistency guarantees of a
single-machine system, despite having dozens or even hundreds of nodes that are working behind the scenes.
For a detailed look at the mechanics behind NewSQL databases, see the DASH: Four Properties of Kubernetes-
Native Databases article.
Data migration to the cloud

One of the more time-consuming tasks is migrating data from one data platform to another. The Azure Data
Migration Service can help expedite such efforts. It can migrate data from several external database sources into
Azure Data platforms with minimal downtime. Target platforms include the following services:
Azure SQL Database
Azure Database for MySQL
Azure Database for MariaDB
Azure Database for PostgreSQL
CosmosDB
The service provides recommendations to guide you through the changes required to execute a migration, both
small or large.
PR EVIO U S NE XT
Caching in a cloud-native app
The benefits of caching are well understood. The technique works by temporarily copying frequently accessed data
from a backend data store to fast storage that's located closer to the application. Caching is often implemented
where...
Data remains relatively static.
Data access is slow, especially compared to the speed of the cache.
Data is subject to high levels of contention.
Why?
As discussed in the Microsoft caching guidance, caching can increase performance, scalability, and availability for
individual microservices and the system as a whole. It reduces the latency and contention of handling large
volumes of concurrent requests to a data store. As data volume and the number of users increase, the greater the
benefits of caching become.
Caching is most effective when a client repeatedly reads data that is immutable or that changes infrequently.
Examples include reference information such as product and pricing information, or shared static resources that
are costly to construct.
While microservices should be stateless, a distributed cache can support concurrent access to session state data
when absolutely required.
Also consider caching to avoid repetitive computations. If an operation transforms data or performs a complicated
calculation, cache the result for subsequent requests.
Caching architecture
Cloud native applications typically implement a distributed caching architecture. The cache is hosted as a cloud-
based backing service, separate from the microservices. Figure 5-15 shows the architecture.
Figure 5-15 : Caching in a cloud native app

In the previous figure, note how the cache is independent of and shared by the microservices. In this scenario, the
cache is invoked by the API Gateway. As discussed in chapter 4, the gateway serves as a front end for all incoming
requests. The distributed cache increases system responsiveness by returning cached data whenever possible.
Additionally, separating the cache from the services allows the cache to scale up or out independently to meet
increased traffic demands.
The previous figure presents a common caching pattern known as the cache-aside pattern. For an incoming
request, you first query the cache (step #1) for a response. If found, the data is returned immediately. If the data
doesn't exist in the cache (known as a cache miss), it's retrieved from a local database in a downstream service
(step #2). It's then written to the cache for future requests (step #3), and returned to the caller. Care must be taken
to periodically evict cached data so that the system remains timely and consistent.
As a shared cache grows, it might prove beneficial to partition its data across multiple nodes. Doing so can help
minimize contention and improve scalability. Many Caching services support the ability to dynamically add and
remove nodes and rebalance data across partitions. This approach typically involves clustering. Clustering exposes
a collection of federated nodes as a seamless, single cache. Internally, however, the data is dispersed across the
nodes following a predefined distribution strategy that balances the load evenly.
Azure Cache for Redis

Azure Cache for Redis is a secure data caching and messaging broker service, fully managed by Microsoft.
Consumed as a Platform as a Service (PaaS) offering, it provides high throughput and low-latency access to data.
The service is accessible to any application within or outside of Azure.
The Azure Cache for Redis service manages access to open-source Redis servers hosted across Azure data centers.
The service acts as a facade providing management, access control, and security. The service natively supports a
rich set of data structures, including strings, hashes, lists, and sets. If your application already uses Redis, it will
work as-is with Azure Cache for Redis.
Azure Cache for Redis is more than a simple cache server. It can support a number of scenarios to enhance a
microservices architecture:
An in-memory data store
A distributed non-relational database
A message broker
A configuration or discovery server
For advanced scenarios, a copy of the cached data can be persisted to disk. If a catastrophic event disables both the
primary and replica caches, the cache is reconstructed from the most recent snapshot.
Azure Redis Cache is available across a number of predefined configurations and pricing tiers. The Premium tier
features many enterprise-level features such as clustering, data persistence, geo-replication, and virtual-network
isolation.
PR EVIO U S NE XT
Elasticsearch in a cloud-native app
Elasticsearch is a distributed search and analytics system that enables complex search capabilities across diverse
types of data. It's open source and widely popular. Consider how the following companies integrate Elasticsearch
into their application:
Wikipedia for full-text and incremental (search as you type) searching.
GitHub to index and expose over 8 million code repositories.
Docker for making its container library discoverable.
Elasticsearch is built on top of the Apache Lucene full-text search engine. Lucene provides high-performance
document indexing and querying. It indexes data with an inverted indexing scheme – instead of mapping pages to
keywords, it maps keywords to pages just like a glossary at the end of a book. Lucene has powerful query syntax
capabilities and can query data by:
Term (a full word)
Prefix (starts-with word)
Wildcard (using "*" or "?" filters)
Phrase (a sequence of text in a document)
Boolean value (complex searches combining queries)
While Lucene provides low-level plumbing for searching, Elasticsearch provides the server that sits on top of
Lucene. Elasticsearch adds higher-level functionality to simplify working Lucene, including a RESTful API to access
Lucene's indexing and searching functionality. It also provides a distributed infrastructure capable of massive
scalability, fault tolerance, and high availability.
For larger cloud-native applications with complex search requirements, Elasticsearch is available as managed
service in Azure. The Microsoft Azure Marketplace features preconfigured templates which developers can use to
deploy an Elasticsearch cluster on Azure.
From the Microsoft Azure Marketplace, developers can use preconfigured templates built to quickly deploy an
Elasticsearch cluster on Azure. Using the Azure-managed offering, you can deploy up to 50 data nodes, 20
coordinating nodes, and three dedicated master nodes.
Summary
This chapter presented a detailed look at data in cloud-native systems. We started by contrasting data storage in
monolithic applications with data storage patterns in cloud-native systems. We looked at data patterns
implemented in cloud-native systems, including cross-service queries, distributed transactions, and patterns to
deal with high-volume systems. We contrasted SQL with NoSQL data. We looked at data storage options available
in Azure that include both Microsoft-centric and open-source options. Finally, we discussed caching and
Elasticsearch in a cloud-native application.
References
Command and Query Responsibility Segregation (CQRS) pattern
Event Sourcing pattern
Why isn't RDBMS Partition Tolerant in CAP Theorem and why is it Available?
Materialized View
All you really need to know about open source databases
Compensating Transaction pattern
Saga Pattern
Saga Patterns | How to implement business transactions using microservices
Compensating Transaction pattern
Getting Behind the 9-Ball: Cosmos DB Consistency Levels Explained
Exploring the different types of NoSQL Databases Part II
On RDBMS, NoSQL and NewSQL databases. Interview with John Ryan
SQL vs NoSQL vs NewSQL: The Full Comparison
DASH: Four Properties of Kubernetes-Native Databases
CockroachDB
TiDB
YugabyteDB
Vitess
Elasticsearch: The Definitive Guide
Introduction to Apache Lucene
PR EVIO U S NE XT
Cloud-native resiliency
Resiliency is the ability of your system to react to failure and still remain functional. It's not about avoiding failure,
but accepting failure and constructing your cloud-native services to respond to it. You want to return to a fully
functioning state quickly as possible.
Unlike traditional monolithic applications, where everything runs together in a single process, cloud-native
systems embrace a distributed architecture as shown in Figure 6-1:
Figure 6-1. Distributed cloud-native environment

In the previous figure, each microservice and cloud-based backing service execute in a separate process, across
server infrastructure, communicating via network-based calls.
Operating in this environment, a service must be sensitive to many different challenges:
Unexpected network latency - the time for a service request to travel to the receiver and back.
Transient faults - short-lived network connectivity errors.
Blockage by a long-running synchronous operation.
A host process that has crashed and is being restarted or moved.
An overloaded microservice that can't respond for a short time.
An in-flight orchestrator operation such as a rolling upgrade or moving a service from one node to another.
Hardware failures.
Cloud platforms can detect and mitigate many of these infrastructure issues. It may restart, scale out, and even
redistribute your service to a different node. However, to take full advantage of this built-in protection, you must
design your services to react to it and thrive in this dynamic environment.
In the following sections, we'll explore defensive techniques that your service and managed cloud resources can
leverage to minimize downtime and disruption.
PR EVIO U S NE XT
Application resiliency patterns
The first line of defense is application resiliency.

While you could invest considerable time writing your own resiliency framework, such products already exist. Polly
is a comprehensive .NET resilience and transient-fault-handling library that allows developers to express resiliency
policies in a fluent and thread-safe manner. Polly targets applications built with either the .NET Framework or .NET
Core. The following table describes the resiliency features, called policies , available in the Polly Library. They can
be applied individually or grouped together.
P O L IC Y EXP ERIEN C E
Retry Configures retry operations on designated operations.
Circuit Breaker Blocks requested operations for a predefined period when

faults exceed a configured threshold
Timeout Places limit on the duration for which a caller can wait for a
response.
Bulkhead Constrains actions to fixed-size resource pool to prevent

failing calls from swamping a resource.
Cache Stores responses automatically.
Fallback Defines structured behavior upon a failure.
Note how in the previous figure the resiliency policies apply to request messages, whether coming from an
external client or back-end service. The goal is to compensate the request for a service that might be momentarily
unavailable. These short-lived interruptions typically manifest themselves with the HTTP status codes shown in the
following table.
H T T P STAT US C O DE C A USE
404 Not Found
408 Request timeout
429 Too many requests (you've most likely been throttled)
502 Bad gateway
503 Service unavailable
504 Gateway timeout
Question: Would you retry an HTTP Status Code of 403 - Forbidden? No. Here, the system is functioning properly,
but informing the caller that they aren't authorized to perform the requested operation. Care must be taken to
retry only those operations caused by failures.
As recommended in Chapter 1, Microsoft developers constructing cloud-native applications should target the .NET
Core platform. Version 2.1 introduced the HTTPClientFactory library for creating HTTP Client instances for
interacting with URL-based resources. Superseding the original HTTPClient class, the factory class supports many
enhanced features, one of which is tight integration with the Polly resiliency library. With it, you can easily define
resiliency policies in the application Startup class to handle partial failures and connectivity issues.
Next, let's expand on retry and circuit breaker patterns.
Retry pattern
In a distributed cloud-native environment, calls to services and cloud resources can fail because of transient (short-
lived) failures, which typically correct themselves after a brief period of time. Implementing a retry strategy helps a
cloud-native service mitigate these scenarios.
The Retry pattern enables a service to retry a failed request operation a (configurable) number of times with an
exponentially increasing wait time. Figure 6-2 shows a retry in action.
Figure 6-2 . Retry pattern in action

In the previous figure, a retry pattern has been implemented for a request operation. It's configured to allow up to
four retries before failing with a backoff interval (wait time) starting at two seconds, which exponentially doubles
for each subsequent attempt.
The first invocation fails and returns an HTTP status code of 500. The application waits for two seconds and
retries the call.
The second invocation also fails and returns an HTTP status code of 500. The application now doubles the
backoff interval to four seconds and retries the call.
Finally, the third call succeeds.
In this scenario, the retry operation would have attempted up to four retries while doubling the backoff
duration before failing the call.
Had the 4th retry attempt failed, a fallback policy would be invoked to gracefully handle the problem.
It's important to increase the backoff period before retrying the call to allow the service time to self-correct. It's a
best practice to implement an exponentially increasing backoff (doubling the period on each retry) to allow
adequate correction time.
Circuit breaker pattern

While the retry pattern can help salvage a request entangled in a partial failure, there are situations where failures
can be caused by unanticipated events that will require longer periods of time to resolve. These faults can range in
severity from a partial loss of connectivity to the complete failure of a service. In these situations, it's pointless for
an application to continually retry an operation that is unlikely to succeed.
To make things worse, executing continual retry operations on a non-responsive service can move you into a self-
imposed denial of service scenario where you flood your service with continual calls exhausting resources such as
memory, threads and database connections, causing failure in unrelated parts of the system that use the same
resources.
In these situations, it would be preferable for the operation to fail immediately and only attempt to invoke the
service if it's likely to succeed.
The Circuit Breaker pattern can prevent an application from repeatedly trying to execute an operation that's likely
to fail. After a pre-defined number of failed calls, it blocks all traffic to the service. Periodically, it will allow a trial
call to determine whether the fault has resolved. Figure 6-3 shows the Circuit Breaker pattern in action.
Figure 6-3 . Circuit breaker pattern in action

In the previous figure, a Circuit Breaker pattern has been added to the original retry pattern. Note how after 100
failed requests, the circuit breakers opens and no longer allows calls to the service. The CheckCircuit value, set at
30 seconds, specifies how often the library allows one request to proceed to the service. If that call succeeds, the
circuit closes and the service is once again available to traffic.
Keep in mind that the intent of the Circuit Breaker pattern is different than that of the Retry pattern. The Retry
pattern enables an application to retry an operation in the expectation that it will succeed. The Circuit Breaker
pattern prevents an application from doing an operation that is likely to fail. Typically, an application will combine
these two patterns by using the Retry pattern to invoke an operation through a circuit breaker.
Testing for resiliency

Testing for resiliency cannot always be done the same way that you test application functionality (by running unit
tests, integration tests and so on). Instead, you must test how the end-to-end workload performs under failure
conditions which only occur intermittently. For example: inject failures by crashing processes, expired certificates,
make dependent services unavailable etc. Frameworks like chaos-monkey can be used for such chaos testing.
Application resiliency is a must for handling problematic requested operations. But, it's only half of the story. Next,
we cover resiliency features available in the Azure cloud.
PR EVIO U S NE XT
Azure platform resiliency
Building a reliable application in the cloud is different from traditional on-premises application development. While
historically you purchased higher-end hardware to scale up, in a cloud environment you scale out. Instead of trying
to prevent failures, the goal is to minimize their effects and keep the system stable.
That said, reliable cloud applications display distinct characteristics:
They're resilient, recover gracefully from problems, and continue to function.
They're highly available (HA) and run as designed in a healthy state with no significant downtime.
Understanding how these characteristics work together - and how they affect cost - is essential to building a
reliable cloud-native application. We'll next look at ways that you can build resiliency and availability into your
cloud-native applications leveraging features from the Azure cloud.
Design with resiliency

We've said resiliency enables your application to react to failure and still remain functional. The whitepaper,
Resilience in Azure whitepaper, provides guidance for achieving resilience in the Azure platform. Here are some
key recommendations:
Hardware failure. Build redundancy into the application by deploying components across different fault
domains. For example, ensure that Azure VMs are placed in different racks by using Availability Sets.
Datacenter failure. Build redundancy into the application with fault isolation zones across datacenters. For
example, ensure that Azure VMs are placed in different fault-isolated datacenters by using Azure Availability
Zones.
Regional failure. Replicate the data and components into another region so that applications can be quickly
recovered. For example, use Azure Site Recovery to replicate Azure VMs to another Azure region.
Heavy load. Load balance across instances to handle spikes in usage. For example, put two or more Azure
VMs behind a load balancer to distribute traffic to all VMs.
Accidental data deletion or corruption. Back up data so it can be restored if there’s any deletion or
corruption. For example, use Azure Backup to periodically back up your Azure VMs.
Design with redundancy

Failures vary in scope of impact. A hardware failure, such as a failed disk, can affect a single node in a cluster. A
failed network switch could affect an entire server rack. Less common failures, such as loss of power, could disrupt
a whole datacenter. Rarely, an entire region becomes unavailable.
Redundancy is one way to provide application resilience. The exact level of redundancy needed depends upon your
business requirements and will affect both the cost and complexity of your system. For example, a multi-region
deployment is more expensive and more complex to manage than a single-region deployment. You'll need
operational procedures to manage failover and failback. The additional cost and complexity might be justified for
some business scenarios, but not others.
To architect redundancy, you need to identify the critical paths in your application, and then determine if there's
redundancy at each point in the path? If a subsystem should fail, will the application fail over to something else?
Finally, you need a clear understanding of those features built into the Azure cloud platform that you can leverage
to meet your redundancy requirements. Here are recommendations for architecting redundancy:
Deploy multiple instances of services. If your application depends on a single instance of a service, it creates
a single point of failure. Provisioning multiple instances improves both resiliency and scalability. When
hosting in Azure Kubernetes Service, you can declaratively configure redundant instances (replica sets) in
the Kubernetes manifest file. The replica count value can be managed programmatically, in the portal, or
through autoscaling features.
Leveraging a load balancer. Load-balancing distributes your application's requests to healthy service
instances and automatically removes unhealthy instances from rotation. When deploying to Kubernetes,
load balancing can be specified in the Kubernetes manifest file in the Services section.
Plan for multiregion deployment. If you deploy your application to a single region, and that region becomes
unavailable, your application will also become unavailable. This may be unacceptable under the terms of
your application's service level agreements. Instead, consider deploying your application and its services
across multiple regions. For example, an Azure Kubernetes Service (AKS) cluster is deployed to a single
region. To protect your system from a regional failure, you might deploy your application to multiple AKS
clusters across different regions and use the Paired Regions feature to coordinate platform updates and
prioritize recovery efforts.
Enable geo-replication. Geo-replication for services such as Azure SQL Database and Cosmos DB will create
secondary replicas of your data across multiple regions. While both services will automatically replicate data
within the same region, geo-replication protects you against a regional outage by enabling you to fail over
to a secondary region. Another best practice for geo-replication centers around storing container images. To
deploy a service in AKS, you need to store and pull the image from a repository. Azure Container Registry
integrates with AKS and can securely store container images. To improve performance and availability,
consider geo-replicating your images to a registry in each region where you have an AKS cluster. Each AKS
cluster then pulls container images from the local container registry in its region as shown in Figure 6-4:
Figure 6-4 . Replicated resources across regions

Implement a DNS traffic load balancer. Azure Traffic Manager provides high-availability for critical applications
by load-balancing at the DNS level. It can route traffic to different regions based on geography, cluster response
time, and even application endpoint health. For example, Azure Traffic Manager can direct customers to the
closest AKS cluster and application instance. If you have multiple AKS clusters in different regions, use Traffic
Manager to control how traffic flows to the applications that run in each cluster. Figure 6-5 shows this scenario.
Figure 6-5 . AKS and Azure Traffic Manager
Design for scalability

The cloud thrives on scaling. The ability to increase/decrease system resources to address increasing/decreasing
system load is a key tenet of the Azure cloud. But, to effectively scale an application, you need an understanding of
the scaling features of each Azure service that you include in your application. Here are recommendations for
effectively implementing scaling in your system.
Design for scaling. An application must be designed for scaling. To start, services should be stateless so that
requests can be routed to any instance. Having stateless services also means that adding or removing an
instance doesn't adversely impact current users.
Partition workloads. Decomposing domains into independent, self-contained microservices enable each
service to scale independently of others. Typically, services will have different scalability needs and
requirements. Partitioning enables you to scale only what needs to be scaled without the unnecessary cost
of scaling an entire application.
Favor scale-out. Cloud-based applications favor scaling out resources as opposed to scaling up. Scaling out
(also known as horizontal scaling) involves adding more service resources to an existing system to meet
and share a desired level of performance. Scaling up (also known as vertical scaling) involves replacing
existing resources with more powerful hardware (more disk, memory, and processing cores). Scaling out
can be invoked automatically with the autoscaling features available in some Azure cloud resources. Scaling
out across multiple resources also adds redundancy to the overall system. Finally scaling up a single
resource is typically more expensive than scaling out across many smaller resources. Figure 6-6 shows the
two approaches:
Figure 6-6. Scale up versus scale out
Scale proportionally. When scaling a service, think in terms of resource sets. If you were to dramatically
scale out a specific service, what impact would that have on back-end data stores, caches and dependent
services? Some resources such as Cosmos DB can scale out proportionally, while many others can't. You
want to ensure that you don't scale out a resource to a point where it will exhaust other associated
resources.
Avoid affinity. A best practice is to ensure a node doesn't require local affinity, often referred to as a sticky
session. A request should be able to route to any instance. If you need to persist state, it should be saved to a
distributed cache, such as Azure Redis cache.
Take advantage of platform autoscaling features. Use built-in autoscaling features whenever possible, rather
than custom or third-party mechanisms. Where possible, use scheduled scaling rules to ensure that
resources are available without a startup delay, but add reactive autoscaling to the rules as appropriate, to
cope with unexpected changes in demand. For more information, see Autoscaling guidance.
Scale out aggressively. A final practice would be to scale out aggressively so that you can quickly meet
immediate spikes in traffic without losing business. And, then scale in (that is, remove unneeded instances)
conservatively to keep the system stable. A simple way to implement this is to set the cool down period,
which is the time to wait between scaling operations, to five minutes for adding resources and up to 15
minutes for removing instances.
Built-in retry in services

We encouraged the best practice of implementing programmatic retry operations in an earlier section. Keep in
mind that many Azure services and their corresponding client SDKs also include retry mechanisms. The following
list summarizes retry features in the many of the Azure services that are discussed in this book:
Azure Cosmos DB. The DocumentClient class from the client API automatically retires failed attempts. The
number of retries and maximum wait time are configurable. Exceptions thrown by the client API are either
requests that exceed the retry policy or non-transient errors.
Azure Redis Cache. The Redis StackExchange client uses a connection manager class that includes retries on
failed attempts. The number of retries, specific retry policy and wait time are all configurable.
Azure Service Bus. The Service Bus client exposes a RetryPolicy class that can be configured with a back-off
interval, retry count, and TerminationTimeBuffer, which specifies the maximum time an operation can take.
The default policy is nine maximum retry attempts with a 30-second backoff period between attempts.
Azure SQL Database. Retry support is provided when using the Entity Framework Core library.
Azure Storage. The storage client library support retry operations. The strategies vary across Azure storage
tables, blobs, and queues. As well, alternate retries switch between primary and secondary storage services
locations when the geo-redundancy feature is enabled.
Azure Event Hubs. The Event Hub client library features a RetryPolicy property, which includes a
configurable exponential backoff feature.
PR EVIO U S NE XT
Resilient communications
Throughout this book, we've embraced a microservice-based architectural approach. While such an architecture
provides important benefits, it presents many challenges:
Out-of-process network communication. Each microservice communicates over a network protocol that
introduces network congestion, latency, and transient faults.
Service discovery. How do microservices discover and communicate with each other when running across a
cluster of machines with their own IP addresses and ports?
Resiliency. How do you manage short-lived failures and keep the system stable?
Load balancing. How does inbound traffic get distributed across multiple instances of a microservice?
Security. How are security concerns such as transport-level encryption and certificate management
enforced?
Distributed Monitoring. - How do you correlate and capture traceability and monitoring for a single request
across multiple consuming microservices?
You can address these concerns with different libraries and frameworks, but the implementation can be expensive,
complex, and time-consuming. You also end up with infrastructure concerns coupled to business logic.
Service mesh
A better approach is an evolving technology entitled Service Mesh. A service mesh is a configurable infrastructure
layer with built-in capabilities to handle service communication and the other challenges mentioned above. It
decouples these concerns by moving them into a service proxy. The proxy is deployed into a separate process
(called a sidecar) to provide isolation from business code. However, the sidecar is linked to the service - it's created
with it and shares its lifecycle. Figure 6-7 shows this scenario.
Figure 6-7 . Service mesh with a side car

In the previous figure, note how the proxy intercepts and manages communication among the microservices and
the cluster.
A service mesh is logically split into two disparate components: A data plane and control plane. Figure 6-8 shows
these components and their responsibilities.
Figure 6-8. Service mesh control and data plane

Once configured, a service mesh is highly functional. It can retrieve a corresponding pool of instances from a
service discovery endpoint. The mesh can then send a request to a specific instance, recording the latency and
response type of the result. A mesh can choose the instance most likely to return a fast response based on many
factors, including its observed latency for recent requests.
If an instance is unresponsive or fails, the mesh will retry the request on another instance. If it returns errors, a
mesh will evict the instance from the load-balancing pool and restate it after it heals. If a request times out, a mesh
can fail and then retry the request. A mesh captures and emits metrics and distributed tracing to a centralized
metrics system.
Istio and Envoy

While a few service mesh options currently exist, Istio is the most popular at the time of this writing. Istio is a joint
venture from IBM, Google, and Lyft. It's an open-source offering that can be integrated into a new or existing
distributed application. The technology provides a consistent and complete solution to secure, connect, and
monitor microservices. Its features include:
Secure service-to-service communication in a cluster with strong identity-based authentication and
authorization.
Automatic load balancing for HTTP, gRPC, WebSocket, and TCP traffic.
Fine-grained control of traffic behavior with rich routing rules, retries, failovers, and fault injection.
A pluggable policy layer and configuration API supporting access controls, rate limits, and quotas.
Automatic metrics, logs, and traces for all traffic within a cluster, including cluster ingress and egress.
A key component for an Istio implementation is a proxy service entitled the Envoy proxy. It runs alongside each
service and provides a platform-agnostic foundation for the following features:
Dynamic service discovery.
Load balancing.
TLS termination.
HTTP and gRPC proxies.
Circuit breaker resiliency.
Health checks.
Rolling updates with canary deployments.
As previously discussed, Envoy is deployed as a sidecar to each microservice in the cluster.
Integration with Azure Kubernetes Services

The Azure cloud embraces Istio and provides direct support for it within Azure Kubernetes Services. The following
links can help you get started:
Installing Istio in AKS
Using AKS and Istio
References
Polly
Retry pattern
Circuit Breaker pattern
Resilience in Azure whitepaper
network latency
Redundancy
geo-replication
Azure Traffic Manager
Autoscaling guidance
Istio
Envoy proxy
PR EVIO U S NE XT
Microservices and cloud-native applications go hand in hand with good DevOps practices. DevOps is many things
to many people but perhaps one of the better definitions comes from cloud advocate and DevOps evangelist
Donovan Brown:
"DevOps is the union of people, process, and products to enable continuous delivery of value to our end users."
Unfortunately, with terse definitions, there's always room to say more things. One of the key components of
DevOps is ensuring that the applications running in production are functioning properly and efficiently. To gauge
the health of the application in production, it's necessary to monitor the various logs and metrics being produced
from the servers, hosts, and the application proper. The number of different services running in support of a cloud-
native application makes monitoring the health of individual components and the application as a whole a critical
challenge.
PR EVIO U S NE XT
Observability patterns
Just as patterns have been developed to aid in the layout of code in applications, there are patterns for operating
applications in a reliable way. Three useful patterns in maintaining applications have emerged: logging ,
monitoring , and aler ts .
When to use logging

No matter how careful we are, applications almost always behave in unexpected ways in production. When users
report problems with an application, it's extremely useful to be able to see what was going on with the app when
the problem occurred. One of the most tried and true ways of capturing information about what an application is
doing while it's running is to have the application write down what it's doing. This process is known as logging. Any
time failures or problems occur in production, the goal should be to reproduce the conditions under which the
failures occurred, in a non-production environment. Having good logging in place provides a roadmap for
developers to follow in order to duplicate problems in an environment that can be tested and experimented with.
Challenges when logging with cloud-native applications
In traditional applications, log files are typically stored on the local machine. In fact, on Unix-like operating systems,
there's a folder structure defined to hold any logs, typically under /var/log .
Figure 7-1 . Logging to a file in a monolithic app.

The usefulness of logging to a flat file on a single machine is vastly reduced in a cloud environment. Applications
producing logs may not have access to the local disk or the local disk may be highly transient as containers are
shuffled around physical machines. Even simple scaling up of monolithic applications across multiple nodes can
make it challenging to locate the appropriate file-based log file.
Figure 7-2 . Logging to files in a scaled monolithic app.
Cloud-native applications developed using a microservices architecture also pose some challenges for file-based
loggers. User requests may now span multiple services that are run on different machines and may include
serverless functions with no access to a local file system at all. It would be very challenging to correlate the logs
from a user or a session across these many services and machines.
Figure 7-3 . Logging to local files in a microservices app.

Finally, the number of users in some cloud-native applications is high. Imagine that each user generates a hundred
lines of log messages when they log into an application. In isolation, that is manageable, but multiply that over
100,000 users and the volume of logs becomes large enough that specialized tools are needed to support effective
use of the logs.
Logging in cloud-native applications
Every programming language has tooling that permits writing logs, and typically the overhead for writing these
logs is low. Many of the logging libraries provide logging different kinds of criticalities, which can be tuned at run
time. For instance, the Serilog library is a popular structured logging library for .NET that provides the following
logging levels:
Verbose
Debug
Information
Warning
Error
Fatal
These different log levels provide granularity in logging. When the application is functioning properly in
production, it may be configured to only log important messages. When the application is misbehaving, then the
log level can be increased so more verbose logs are gathered. This balances performance against ease of
debugging.
The high performance of logging tools and the tunability of verbosity should encourage developers to log
frequently. Many favor a pattern of logging the entry and exit of each method. This approach may sound like
overkill, but it's infrequent that developers will wish for less logging. In fact, it's not uncommon to perform
deployments for the sole purpose of adding logging around a problematic method. Err on the side of too much
logging and not on too little. Some tools can be used to automatically provide this kind of logging.
Because of the challenges associated with using file-based logs in cloud-native apps, centralized logs are preferred.
Logs are collected by the applications and shipped to a central logging application which indexes and stores the
logs. This class of system can ingest tens of gigabytes of logs every day.
It's also helpful to follow some standard practices when building logging that spans many services. For instance,
generating a correlation ID at the start of a lengthy interaction, and then logging it in each message that is related
to that interaction, makes it easier to search for all related messages. One need only find a single message and
extract the correlation ID to find all the related messages. Another example is ensuring that the log format is the
same for every service, whatever the language or logging library it uses. This standardization makes reading logs
much easier. Figure 7-4 demonstrates how a microservices architecture can leverage centralized logging as part of
its workflow.
Figure 7-4 . Logs from various sources are ingested into a centralized log store.
Challenges with detecting and responding to potential app health
issues
Some applications aren't mission critical. Maybe they're only used internally, and when a problem occurs, the user
can contact the team responsible and the application can be restarted. However, customers often have higher
expectations for the applications they consume. You should know when problems occur with your application
before users do, or before users notify you. Otherwise, the first you know about a problem may be when you
notice an angry deluge of social media posts deriding your application or even your organization.
Some scenarios you may need to consider include:
One service in your application keeps failing and restarting, resulting in intermittent slow responses.
At some times of the day, your application's response time is slow.
After a recent deployment, load on the database has tripled.
Implemented properly, monitoring can let you know about conditions that will lead to problems, letting you
address underlying conditions before they result in any significant user impact.
Monitoring cloud-native apps
Some centralized logging systems take on an additional role of collecting telemetry outside of pure logs. They can
collect metrics, such as time to run a database query, average response time from a web server, and even CPU load
averages and memory pressure as reported by the operating system. In conjunction with the logs, these systems
can provide a holistic view of the health of nodes in the system and the application as a whole.
The metric-gathering capabilities of the monitoring tools can also be fed manually from within the application.
Business flows that are of particular interest such as new users signing up or orders being placed, may be
instrumented such that they increment a counter in the central monitoring system. This unlocks the monitoring
tools to not only monitor the health of the application but the health of the business.
Queries can be constructed in the log aggregation tools to look for certain statistics or patterns, which can then be
displayed in graphical form, on custom dashboards. Frequently, teams will invest in large, wall-mounted displays
that rotate through the statistics related to an application. This way, it's simple to see the problems as they occur.
Cloud-native monitoring tools provide real-time telemetry and insight into apps regardless of whether they're
single-process monolithic applications or distributed microservice architectures. They include tools that allow
collection of data from the app as well as tools for querying and displaying information about the app's health.
Challenges with reacting to critical problems in cloud-native apps

If you need to react to problems with your application, you need some way to alert the right personnel. This is the
third cloud-native application observability pattern and depends on logging and monitoring. Your application
needs to have logging in place to allow problems to be diagnosed, and in some cases to feed into monitoring tools.
It needs monitoring to aggregate application metrics and health data in one place. Once this has been established,
rules can be created that will trigger alerts when certain metrics fall outside of acceptable levels.
Generally, alerts are layered on top of monitoring such that certain conditions trigger appropriate alerts to notify
team members of urgent problems. Some scenarios that may require alerts include:
One of your application's services is not responding after 1 minute of downtime.
Your application is returning unsuccessful HTTP responses to more than 1% of requests.
Your application's average response time for key endpoints exceeds 2000 ms.
Alerts in cloud-native apps
You can craft queries against the monitoring tools to look for known failure conditions. For instance, queries could
search through the incoming logs for indications of HTTP status code 500, which indicates a problem on a web
server. As soon as one of these is detected, then an e-mail or an SMS could be sent to the owner of the originating
service who can begin to investigate.
Typically, though, a single 500 error isn't enough to determine that a problem has occurred. It could mean that a
user mistyped their password or entered some malformed data. The alert queries can be crafted to only fire when
a larger than average number of 500 errors are detected.
One of the most damaging patterns in alerting is to fire too many alerts for humans to investigate. Service owners
will rapidly become desensitized to errors that they've previously investigated and found to be benign. Then, when
true errors occur, they'll be lost in the noise of hundreds of false positives. The parable of the Boy Who Cried Wolf
is frequently told to children to warn them of this very danger. It's important to ensure that the alerts that do fire
are indicative of a real problem.
PR EVIO U S NE XT
Logging with Elastic Stack
There are many good centralized logging tools and they vary in cost from being free, open-source tools, to more
expensive options. In many cases, the free tools are as good as or better than the paid offerings. One such tool is a
combination of three open-source components: Elastic search, Logstash, and Kibana.
Collectively these tools are known as the Elastic Stack or ELK stack.
Elastic Stack
The Elastic Stack is a powerful option for gathering information from a Kubernetes cluster. Kubernetes supports
sending logs to an Elasticsearch endpoint, and for the most part, all you need to get started is to set the
environment variables as shown in Figure 7-5:
KUBE_LOGGING_DESTINATION=elasticsearch
KUBE_ENABLE_NODE_LOGGING=true
Figure 7-5 . Configuration variables for Kubernetes

This will install Elasticsearch on the cluster and target sending all the cluster logs to it.
Figure 7-6 . An example of a Kibana dashboard showing the results of a query against logs that are ingested from
Kubernetes
What are the advantages of Elastic Stack?

Elastic Stack provides centralized logging in a low-cost, scalable, cloud-friendly manner. Its user interface
streamlines data analysis so you can spend your time gleaning insights from your data instead of fighting with a
clunky interface. It supports a wide variety of inputs so as your distributed application spans more and different
kinds of services, you can expect to continue to be able to feed log and metric data into the system. The Elastic
Stack also supports fast searches even across large data sets, making it possible even for large applications to log
detailed data and still be able to have visibility into it in a performant fashion.
Logstash
The first component is Logstash. This tool is used to gather log information from a large variety of different
sources. For instance, Logstash can read logs from disk and also receive messages from logging libraries like
Serilog. Logstash can do some basic filtering and expansion on the logs as they arrive. For instance, if your logs
contain IP addresses then Logstash may be configured to do a geographical lookup and obtain a country or even
city of origin for that message.
Serilog is a logging library for .NET languages, which allows for parameterized logging. Instead of generating a
textual log message that embeds fields, parameters are kept separate. This allows for more intelligent filtering and
searching. A sample Serilog configuration for writing to Logstash appears in Figure 7-7.
var log = new LoggerConfiguration()

.WriteTo.Http("http://localhost:8080")
.CreateLogger();
Figure 7-7 . Serilog config for writing log information directly to logstash over HTTP
Logstash would use a configuration like the one shown in Figure 7-8.
input {
http {
#default host 0.0.0.0:8080
codec => json
}
}
output {
elasticsearch {
hosts => "elasticsearch:9200"
index=>"sales-%{+xxxx.ww}"
}
}
Figure 7-8 . A Logstash configuration for consuming logs from Serilog

For scenarios where extensive log manipulation isn't needed there's an alternative to Logstash known as Beats.
Beats is a family of tools that can gather a wide variety of data from logs to network data and uptime information.
Many applications will use both Logstash and Beats.
Once the logs have been gathered by Logstash, it needs somewhere to put them. While Logstash supports many
different outputs, one of the more exciting ones is Elastic search.
Elastic search
Elastic search is a powerful search engine that can index logs as they arrive. It makes running queries against the
logs quick. Elastic search can handle huge quantities of logs and, in extreme cases, can be scaled out across many
nodes.
Log messages that have been crafted to contain parameters or that have had parameters split from them through
Logstash processing, can be queried directly as Elasticsearch preserves this information.
A query that searches for the top 10 pages visited by [email protected] , appears in Figure 7-9.
"query": {
"match": {
"user": "[email protected]"
}
},
"aggregations": {
"top_10_pages": {
"terms": {
"field": "page",
"size": 10
}
}
}
Figure 7-9 . An Elasticsearch query for finding top 10 pages visited by a user
Visualizing information with Kibana web dashboards

The final component of the stack is Kibana. This tool is used to provide interactive visualizations in a web
dashboard. Dashboards may be crafted even by users who are non-technical. Most data that is resident in the
Elasticsearch index, can be included in the Kibana dashboards. Individual users may have different dashboard
desires and Kibana enables this customization through allowing user-specific dashboards.
Installing Elastic Stack on Azure

The Elastic stack can be installed on Azure in a number of ways. As always, it's possible to provision virtual
machines and install Elastic Stack on them directly. This option is preferred by some experienced users as it offers
the highest degree of customizability. Deploying on infrastructure as a service introduces significant management
overhead forcing those who take that path to take ownership of all the tasks associated with infrastructure as a
service such as securing the machines and keeping up-to-date with patches.
An option with less overhead is to make use of one of the many Docker containers on which the Elastic Stack has
already been configured. These containers can be dropped into an existing Kubernetes cluster and run alongside
application code. The sebp/elk container is a well-documented and tested Elastic Stack container.
Another option is a recently announced ELK-as-a-service offering.
References
Install Elastic Stack on Azure
PR EVIO U S NE XT
Monitoring in Azure Kubernetes Services
The built-in logging in Kubernetes is primitive. However, there are some great options for getting the logs out of
Kubernetes and into a place where they can be properly analyzed. If you need to monitor your AKS clusters,
configuring Elastic Stack for Kubernetes is a great solution.
Azure Monitor for Containers

Azure Monitor for Containers supports consuming logs from not just Kubernetes but also from other orchestration
engines such as DC/OS, Docker Swarm, and Red Hat OpenShift.
Figure 7-10 . Consuming logs from various containers

Prometheus is a popular open source metric monitoring solution. It is part of the Cloud Native Compute
Foundation. Typically, using Prometheus requires managing a Prometheus server with its own store. However,
Azure Monitor for Containers provides direct integration with Prometheus metrics endpoints, so a separate server
is not required.
Log and metric information is gathered not just from the containers running in the cluster but also from the cluster
hosts themselves. It allows correlating log information from the two making it much easier to track down an error.
Installing the log collectors differs on Windows and Linux clusters. But in both cases the log collection is
implemented as a Kubernetes DaemonSet, meaning that the log collector is run as a container on each of the
nodes.
No matter which orchestrator or operating system is running the Azure Monitor daemon, the log information is
forwarded to the same Azure Monitor tools with which users are familiar. This ensures a parallel experience in
environments that mix different log sources such as a hybrid Kubernetes/Azure Functions environment.
Figure 7-11 . A sample dashboard showing logging and metric information from a number of running containers.
Log.Finalize()
Logging is one of the most overlooked and yet most important parts of deploying any application at scale. As the
size and complexity of applications increase, then so does the difficulty of debugging them. Having top quality logs
available makes debugging much easier and moves it from the realm of "nearly impossible" to "a pleasant
experience".
PR EVIO U S NE XT
Azure Monitor
No other cloud provider has as mature of a cloud application monitoring solution as that found in Azure. Azure
Monitor is an umbrella name for a collection of tools designed to provide visibility into the state of your system,
insights into any problems, and optimization of your application.
Figure 7-12 . Azure Monitor, a collection to tools to provide insight into how a cloud-native application is
functioning.
Gathering logs and metrics

The first step in any monitoring solution is to gather as much data as possible. The more data that can be gathered,
the deeper the insights that can be obtained. Instrumenting systems has traditionally been difficult. Simple
Network Management Protocol (SNMP) was the gold standard protocol for collecting machine level information
but it required a great deal of knowledge and configuring. Fortunately, much of this hard work has been eliminated
as the most common metrics are gathered automatically by Azure Monitor.
Application level metrics and events aren't possible to instrument automatically because they're specific to the
application being deployed. In order to gather these metrics, there are SDKs and APIs available to directly report
such information, such as when a customer signs up or completes an order. Exceptions can also be captured and
reported back into Azure Monitor via Application Insights. The SDKs support most every language found in Cloud
Native Applications including Go, Python, JavaScript, and the .NET languages.
The ultimate goal of gathering information about the state of your application is to ensure that your end users
have a good experience. What better way to tell if users are experiencing issues than doing outside-in web tests?
These tests can be as simple as pinging your website from locations around the world or as involved as having
agents log into the site and simulate user actions.
Reporting data
Once the data is gathered, it can be manipulated, summarized, and plotted into charts, which allow users to
instantly see when there are problems. These charts can be gathered into dashboards or into Workbooks, a multi-
page report designed to tell a story about some aspect of the system.
No modern application would be complete without some artificial intelligence or machine learning. To this end,
data can be passed to the various machine learning tools in Azure to allow you to extract trends and information
that would otherwise be hidden.
Application Insights provides a powerful query language called Kusto that can be used to find records, summarize
them, and even plot charts. For instance, this query will locate all the records for the month of November 2007,
group them by state, and plot the top 10 as a pie chart.
StormEvents
| where StartTime >= datetime(2007-11-01) and StartTime < datetime(2007-12-01)
| summarize count() by State
| top 10 by count_
| render piechart
Figure 7-13 . The result of the Application Insights Query.

There is a playground for experimenting with Kusto queries, which is a fantastic place to spend an hour or two.
Reading sample queries can also be instructive.
Dashboards
There are several different dashboard technologies that may be used to surface the information from Azure
Monitor. Perhaps the simplest is to just run queries in Application Insights and plot the data into a chart.
Figure 7-14 . An example of Application Insights charts embedded in the main Azure Dashboard.
These charts can then be embedded in the Azure portal proper through use of the dashboard feature. For users
with more exacting requirements, such as being able to drill down into several tiers of data, Azure Monitor data is
available to Power BI. Power BI is an industry-leading, enterprise class, business intelligence tool that can aggregate
data from many different data sources.
Figure 7-15 . An example Power BI dashboard.
Alerts
Sometimes, having data dashboards is insufficient. If nobody is awake to watch the dashboards, then it can still be
many hours before a problem is addressed, or even detected. To this end, Azure Monitor also provides a top notch
alerting solution. Alerts can be triggered by a wide range of conditions including:
Metric values
Log search queries
Activity Log events
Health of the underlying Azure platform
Tests for web site availability
When triggered, the alerts can perform a wide variety of tasks. On the simple side, the alerts may just send an e-
mail notification to a mailing list or a text message to an individual. More involved alerts might trigger a workflow
in a tool such as PagerDuty, which is aware of who is on call for a particular application. Alerts can trigger actions
in Microsoft Flow unlocking near limitless possibilities for workflows.
As common causes of alerts are identified, the alerts can be enhanced with details about the common causes of the
alerts and the steps to take to resolve them. Highly mature cloud-native application deployments may opt to kick
off self-healing tasks, which perform actions such as removing failing nodes from a scale set or triggering an
autoscaling activity. Eventually it may no longer be necessary to wake up on-call personnel at 2AM to resolve a
live-site issue as the system will be able to adjust itself to compensate or at least limp along until somebody arrives
at work the next morning.
Azure Monitor automatically leverages machine learning to understand the normal operating parameters of
deployed applications. This enables it to detect services that are operating outside of their normal parameters. For
instance, the typical weekday traffic on the site might be 10,000 requests per minute. And then, on a given week,
suddenly the number of requests hits a highly unusual 20,000 requests per minute. Smart Detection will notice this
deviation from the norm and trigger an alert. At the same time, the trend analysis is smart enough to avoid firing
false positives when the traffic load is expected.
References
Azure Monitor
PR EVIO U S NE XT
Identity
Most software applications need to have some knowledge of the user or process that is calling them. The user or
process interacting with an application is known as a security principal, and the process of authenticating and
authorizing these principals is known as identity management, or simply identity. Simple applications may include
all of their identity management within the application, but this approach doesn't scale well with many applications
and many kinds of security principals. Windows supports the use of Active Directory to provide centralized
authentication and authorization.
While this solution is effective within corporate networks, it isn't designed for use by users or applications that are
outside of the AD domain. With the growth of Internet-based applications and the rise of cloud-native apps,
security models have evolved.
In today's cloud-native identity model, architecture is assumed to be distributed. Apps can be deployed anywhere
and may communicate with other apps anywhere. Clients may communicate with these apps from anywhere, and
in fact, clients may consist of any combination of platforms and devices. Cloud-native identity solutions leverage
open standards to achieve secure application access from clients. These clients range from human users on PCs or
phones, to other apps hosted anywhere online, to set-top boxes and IOT devices running any software platform
anywhere in the world.
Modern cloud-native identity solutions typically leverage access tokens that are issued by a secure token
service/server (STS) to a security principal once their identity is determined. The access token, typically a JSON
Web Token (JWT), includes claims about the security principal. These claims will minimally include the user's
identity but may also include additional claims that can be used by applications to determine the level of access to
grant the principal.
Typically, the STS is only responsible for authenticating the principal. Determining their level of access to resources
is left to other parts of the application.
References
Microsoft identity platform
PR EVIO U S NE XT
Authentication and authorization in cloud-native
apps
Authentication is the process of determining the identity of a security principal. Authorization is the act of granting
an authenticated principal permission to perform an action or access a resource. Sometimes authentication is
shortened to AuthN and authorization is shortened to AuthZ . Cloud-native applications need to rely on open
HTTP-based protocols to authenticate security principals since both clients and applications could be running
anywhere in the world on any platform or device. The only common factor is HTTP.
Many organizations still rely on local authentication services like Active Directory Federation Services (ADFS).
While this approach has traditionally served organizations well for on premises authentication needs, cloud-native
applications benefit from systems designed specifically for the cloud. A recent 2019 United Kingdom National
Cyber Security Centre (NCSC) advisory states that "organizations using Azure AD as their primary authentication
source will actually lower their risk compared to ADFS." Some reasons outlined in this analysis include:
Access to full set of Microsoft credential protection technologies.
Most organizations are already relying on Azure AD to some extent.
Double hashing of NTLM hashes ensures compromise won't allow credentials that work in local Active
Directory.
References
Authentication basics
Access tokens and claims
It may be time to ditch your on premises authentication services
PR EVIO U S NE XT
Azure Active Directory
Microsoft Azure Active Directory (Azure AD) offers identity and access management as a service. Customers use it
to configure and maintain who users are, what information to store about them, who can access that information,
who can manage it, and what apps can access it. AAD can authenticate users for applications configured to use it,
providing a single sign-on (SSO) experience. It can be used on its own or be integrated with Windows AD running
on premises.
Azure AD is built for the cloud. It's truly a cloud-native identity solution that uses a REST-based Graph API and
OData syntax for queries, unlike Windows AD, which uses LDAP. On premises Active Directory can sync user
attributes to the cloud using Identity Sync Services, allowing all authentication to take place in the cloud using
Azure AD. Alternately, authentication can be configured via Connect to pass back to local Active Directory via ADFS
to be completed by Windows AD on premises.
Azure AD supports company branded sign-in screens, multi-factory authentication, and cloud-based application
proxies that are used to provide SSO for applications hosted on premises. It offers different kinds of security
reporting and alert capabilities.
References
Microsoft identity platform
PR EVIO U S NE XT
IdentityServer for cloud-native applications
IdentityServer is an open-source authentication server that implements OpenID Connect (OIDC) and OAuth 2.0
standards for ASP.NET Core. It's designed to provide a common way to authenticate requests to all of your
applications, whether they're web, native, mobile, or API endpoints. IdentityServer can be used to implement Single
Sign-On (SSO) for multiple applications and application types. It can be used to authenticate actual users via sign-
in forms and similar user interfaces as well as service-based authentication that typically involves token issuance,
verification, and renewal without any user interface. IdentityServer is designed to be a customizable solution. Each
instance is typically customized to suit an individual organization and/or set of applications' needs.
Common web app scenarios

Typically, applications need to support some or all of the following scenarios:
Human users accessing web applications with a browser.
Human users accessing back-end Web APIs from browser-based apps.
Human users on mobile/native clients accessing back-end Web APIs.
Other applications accessing back-end Web APIs (without an active user or user interface).
Any application may need to interact with other Web APIs, using its own identity or delegating to the user's
identity.
Figure 8-1 . Application types and scenarios.

In each of these scenarios, the exposed functionality needs to be secured against unauthorized use. At a minimum,
this typically requires authenticating the user or principal making a request for a resource. This authentication may
use one of several common protocols such as SAML2p, WS-Fed, or OpenID Connect. Communicating with APIs
typically uses the OAuth2 protocol and its support for security tokens. Separating these critical cross-cutting
security concerns and their implementation details from the applications themselves ensures consistency and
improves security and maintainability. Outsourcing these concerns to a dedicated product like IdentityServer helps
the requirement for every application to solve these problems itself.
IdentityServer provides middleware that runs within an ASP.NET Core application and adds support for OpenID
Connect and OAuth2 (see supported specifications). Organizations would create their own ASP.NET Core app using
IdentityServer middleware to act as the STS for all of their token-based security protocols. The IdentityServer
middleware exposes endpoints to support standard functionality, including:
Authorize (authenticate the end user)
Token (request a token programmatically)
Discovery (metadata about the server)
User Info (get user information with a valid access token)
Device Authorization (used to start device flow authorization)
Introspection (token validation)
Revocation (token revocation)
End Session (trigger single sign-out across all apps)
Getting started
IdentityServer4 is open-source and free to use. You can add it to your applications using its NuGet packages. The
main package is IdentityServer4 that has been downloaded over four million times. The base package doesn't
include any user interface code and only supports in memory configuration. To use it with a database, you'll also
want a data provider like IdentityServer4.EntityFramework that uses Entity Framework Core to store configuration
and operational data for IdentityServer. For user interface, you can copy files from the Quickstart UI repository into
your ASP.NET Core MVC application to add support for sign in and sign out using IdentityServer middleware.
Configuration
IdentityServer supports different kinds of protocols and social authentication providers that can be configured as
part of each custom installation. This is typically done in the ASP.NET Core application's Startup class in the
ConfigureServices method. The configuration involves specifying the supported protocols and the paths to the
servers and endpoints that will be used. Figure 8-2 shows an example configuration taken from the
IdentityServer4 Quickstart UI project:
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddMvc();
// some details omitted

services.AddIdentityServer();
services.AddAuthentication()
.AddGoogle("Google", options =>
{
options.SignInScheme = IdentityServerConstants.ExternalCookieAuthenticationScheme;
options.ClientId = "<insert here>";

options.ClientSecret = "<insert here>";
})
.AddOpenIdConnect("demoidsrv", "IdentityServer", options =>
{
options.SignInScheme = IdentityServerConstants.ExternalCookieAuthenticationScheme;
options.SignOutScheme = IdentityServerConstants.SignoutScheme;
options.Authority = "https://demo.identityserver.io/";
options.ClientId = "implicit";
options.ResponseType = "id_token";
options.SaveTokens = true;
options.CallbackPath = new PathString("/signin-idsrv");
options.SignedOutCallbackPath = new PathString("/signout-callback-idsrv");
options.RemoteSignOutPath = new PathString("/signout-idsrv");
options.TokenValidationParameters = new TokenValidationParameters

{
NameClaimType = "name",
RoleClaimType = "role"
};
});
}
}
Figure 8-2 . Configuring IdentityServer.

IdentityServer also hosts a public demo site that can be used to test various protocols and configurations. It's
located at https://demo.identityserver.io/ and includes information on how to configure its behavior based on the
client_id provided to it.
JavaScript clients
Many cloud-native applications leverage server-side APIs and rich client single page applications (SPAs) on the
front end. IdentityServer ships a JavaScript client ( oidc-client.js ) via NPM that can be added to SPAs to enable
them to use IdentityServer for sign in, sign out, and token-based authentication of web APIs.
References
IdentityServer documentation
Application types
JavaScript OIDC client
PR EVIO U S NE XT
Security
Not a day goes by where the news doesn't contain some story about a company being hacked or somehow losing
their customers' data. Even countries aren't immune to the problems created by treating security as an
afterthought. For years, companies have treated the security of customer data and, in fact, their entire networks as
something of a "nice to have". Windows servers were left unpatched, ancient versions of PHP kept running and
MongoDB databases left wide open to the world.
However, there are starting to be real-world consequences for not maintaining a security mindset when building
and deploying applications. Many companies learned the hard way what can happen when servers and desktops
aren't patched during the 2017 outbreak of NotPetya. The cost of these attacks has easily reached into the billions,
with some estimates putting the losses from this single attack at 10 billion US dollars.
Even governments aren't immune to hacking incidents. The city of Baltimore was held ransom by criminals making
it impossible for citizens to pay their bills or use city services.
There has also been an increase in legislation that mandates certain data protections for personal data. In Europe,
GDPR has been in effect for more than a year and, more recently, California passed their own version called CCDA,
which comes into effect January 1, 2020. The fines under GDPR can be so punishing as to put companies out of
business. Google has already been fined 50 million Euros for violations, but that's just a drop in the bucket
compared with the potential fines.
In short, security is serious business.
PR EVIO U S NE XT
Azure security for cloud-native apps
Cloud-native applications can be both easier and more difficult to secure than traditional applications. On the
downside, you need to secure more smaller applications and dedicate more energy to build out the security
infrastructure. The heterogeneous nature of programming languages and styles in most service deployments also
means you need to pay more attention to security bulletins from many different providers.
On the flip side, smaller services, each with their own data store, limit the scope of an attack. If an attacker
compromises one system, it's probably more difficult for the attacker to make the jump to another system than it is
in a monolithic application. Process boundaries are strong boundaries. Also, if a database backup leaks, then the
damage is more limited, as that database contains only a subset of data and is unlikely to contain personal data.
Threat modeling
No matter if the advantages outweigh the disadvantages of cloud-native applications, the same holistic security
mindset must be followed. Security and secure thinking must be part of every step of the development and
operations story. When planning an application ask questions like:
What would be the impact of this data being lost?
How can we limit the damage from bad data being injected into this service?
Who should have access to this data?
Are there auditing policies in place around the development and release process?
All these questions are part of a process called threat modeling. This process tries to answer the question of what
threats there are to the system, how likely the threats are, and the potential damage from them.
Once the list of threats has been established, you need to decide whether they're worth mitigating. Sometimes a
threat is so unlikely and expensive to plan for that it isn't worth spending energy on it. For instance, some state
level actor could inject changes into the design of a process that is used by millions of devices. Now, instead of
running a certain piece of code in Ring 3, that code is run in Ring 0. This allows an exploit that can bypass the
hypervisor and run the attack code on the bare metal machines, allowing attacks on all the virtual machines that
are running on that hardware.
The altered processors are difficult to detect without a microscope and advanced knowledge of the on silicon
design of that processor. This scenario is unlikely to happen and expensive to mitigate, so probably no threat model
would recommend building exploit protection for it.
More likely threats, such as broken access controls permitting Id incrementing attacks (replacing Id=2 with
Id=3 in the URL) or SQL injection, are more attractive to build protections against. The mitigations for these
threats are quite reasonable to build and prevent embarrassing security holes that smear the company's
reputation.
Principle of least privilege

One of the founding ideas in computer security is the Principle of Least Privilege (POLP). It's actually a foundational
idea in most any form of security be it digital or physical. In short, the principle is that any user or process should
have the smallest number of rights possible to execute its task.
As an example, think of the tellers at a bank: accessing the safe is an uncommon activity. So, the average teller can't
open the safe themselves. To gain access, they need to escalate their request through a bank manager, who
performs additional security checks.
In a computer system, a fantastic example is the rights of a user connecting to a database. In many cases, there's a
single user account used to both build the database structure and run the application. Except in extreme cases, the
account running the application doesn't need the ability to update schema information. There should be several
accounts that provide different levels of privilege. The application should only use the permission level that grants
read and write access to the data in the tables. This kind of protection would eliminate attacks that aimed to drop
database tables or introduce malicious triggers.
Almost every part of building a cloud-native application can benefit from remembering the principle of least
privilege. You can find it at play when setting up firewalls, network security groups, roles, and scopes in Role-based
access control (RBAC).
Penetration testing
As applications become more complicated the number of attack vectors increases at an alarming rate. Threat
modeling is flawed in that it tends to be executed by the same people building the system. In the same way that
many developers have trouble envisioning user interactions and then build unusable user interfaces, most
developers have difficulty seeing every attack vector. It's also possible that the developers building the system
aren't well versed in attack methodologies and miss something crucial.
Penetration testing or "pen testing" involves bringing in external actors to attempt to attack the system. These
attackers may be an external consulting company or other developers with good security knowledge from another
part of the business. They're given carte blanche to attempt to subvert the system. Frequently, they'll find extensive
security holes that need to be patched. Sometimes the attack vector will be something totally unexpected like
exploiting a phishing attack against the CEO.
Azure itself is constantly undergoing attacks from a team of hackers inside Microsoft. Over the years, they've been
the first to find dozens of potentially catastrophic attack vectors, closing them before they can be exploited
externally. The more tempting a target, the more likely that eternal actors will attempt to exploit it and there are a
few targets in the world more tempting than Azure.
Monitoring
Should an attacker attempt to penetrate an application, there should be some warning of it. Frequently, attacks can
be spotted by examining the logs from services. Attacks leave telltale signs that can be spotted before they
succeed. For instance, an attacker attempting to guess a password will make many requests to a login system.
Monitoring around the login system can detect weird patterns that are out of line with the typical access pattern.
This monitoring can be turned into an alert that can, in turn, alert an operations person to activate some sort of
countermeasure. A highly mature monitoring system might even take action based on these deviations proactively
adding rules to block requests or throttle responses.
Securing the build

One place where security is often overlooked is around the build process. Not only should the build run security
checks, such as scanning for insecure code or checked-in credentials, but the build itself should be secure. If the
build server is compromised, then it provides a fantastic vector for introducing arbitrary code into the product.
Imagine that an attacker is looking to steal the passwords of people signing into a web application. They could
introduce a build step that modifies the checked-out code to mirror any login request to another server. The next
time code goes through the build, it's silently updated. The source code vulnerability scanning won't catch this as it
runs before the build. Equally, nobody will catch it in a code review because the build steps live on the build server.
The exploited code will go to production where it can harvest passwords. Probably there's no audit log of the build
process changes, or at least nobody monitoring the audit.
This is a perfect example of a seemingly low value target that can be used to break into the system. Once an
attacker breaches the perimeter of the system, they can start working on finding ways to elevate their permissions
to the point that they can cause real harm anywhere they like.
Building secure code

.NET Framework is already a quite secure framework. It avoids some of the pitfalls of unmanaged code, such as
walking off the ends of arrays. Work is actively done to fix security holes as they're discovered. There's even a bug
bounty program that pays researchers to find issues in the framework and report them instead of exploiting them.
There are many ways to make .NET code more secure. Following guidelines such as the Secure coding guidelines
for .NET article is a reasonable step to take to ensure that the code is secure from the ground up. The OWASP top
10 is another invaluable guide to build secure code.
The build process is a good place to put scanning tools to detect problems in source code before they make it into
production. Most every project has dependencies on some other packages. A tool that can scan for outdated
packages will catch problems in a nightly build. Even when building Docker images, it's useful to check and make
sure that the base image doesn't have known vulnerabilities. Another thing to check is that nobody has accidentally
checked in credentials.
Built-in security
Azure is designed to balance usability and security for the majority of users. Different users are going to have
different security requirements, so they need to fine-tune their approach to cloud security. Microsoft publishes a
great deal of security information in the Trust Center. This resource should be the first stop for those professionals
interested in understanding how the built-in attack mitigation technologies work.
Within the Azure portal, the Azure Advisor is a system that is constantly scanning an environment and making
recommendations. Some of these recommendations are designed to save users money, but others are designed to
identify potentially insecure configurations, such as having a storage container open to the world and not
protected by a Virtual Network.
Azure network infrastructure

In an on-premises deployment environment, a great deal of energy is dedicated to setting up networking. Setting
up routers, switches, and the such is complicated work. Networks allow certain resources to talk to other resources
and prevent access in some cases. A frequent network rule is to restrict access to the production environment from
the development environment on the off chance that a half-developed piece of code runs awry and deletes a swath
of data.
Out of the box, most PaaS Azure resources have only the most basic and permissive networking setup. For
instance, anybody on the Internet can access an app service. New SQL Server instances typically come restricted,
so that external parties can't access them, but the IP address ranges used by Azure itself are permitted through. So,
while the SQL server is protected from external threats, an attacker only needs to set up an Azure bridgehead from
where they can launch attacks against all SQL instances on Azure.
Fortunately, most Azure resources can be placed into an Azure Virtual Network that allows finer grained access
control. Similar to the way that on-premises networks establish private networks that are protected from the wider
world, virtual networks are islands of private IP addresses that are located within the Azure network.
Figure 9-1 . A virtual network in Azure.
In the same way that on-premises networks have a firewall governing access to the network, you can establish a
similar firewall at the boundary of the virtual network. By default, all the resources on a virtual network can still
talk to the Internet. It's only incoming connections that require some form of explicit firewall exception.
With the network established, internal resources like storage accounts can be set up to only allow for access by
resources that are also on the Virtual Network. This firewall provides an extra level of security, should the keys for
that storage account be leaked, attackers wouldn't be able to connect to it to exploit the leaked keys. This is another
example of the principle of least privilege.
The nodes in an Azure Kubernetes cluster can participate in a virtual network just like other resources that are
more native to Azure. This functionality is called Azure Container Networking Interface. In effect, it allocates a
subnet within the virtual network on which virtual machines and container images are allocated.
Continuing down the path of illustrating the principle of least privilege, not every resource within a Virtual
Network needs to talk to every other resource. For instance, in an application that provides a web API over a
storage account and a SQL database, it's unlikely that the database and the storage account need to talk to one
another. Any data sharing between them would go through the web application. So, a network security group
(NSG) could be used to deny traffic between the two services.
A policy of denying communication between resources can be annoying to implement, especially coming from a
background of using Azure without traffic restrictions. On some other clouds, the concept of network security
groups is much more prevalent. For instance, the default policy on AWS is that resources can't communicate
among themselves until enabled by rules in an NSG. While slower to develop this, more restrictive environment
provides a more secure default. Making use of proper DevOps practices, especially using Azure Resource Manager
or Terraform to manage permissions can make controlling the rules easier.
Virtual Networks can also be useful when setting up communication between on-premises and cloud resources. A
virtual private network can be used to seamlessly attach the two networks together. This allows running a virtual
network without any sort of gateway for scenarios where all the users are on-site. There are a number of
technologies that can be used to establish this network. The simplest is to use a site-to-site VPN that can be
established between many routers and Azure. Traffic is encrypted and tunneled over the Internet at the same cost
per byte as any other traffic. For scenarios where more bandwidth or more security is desirable, Azure offers a
service called Express Route that uses a private circuit between an on-premises network and Azure. It's more costly
and difficult to establish but also more secure.
Role-based access control for restricting access to Azure resources
RBAC is a system that provides an identity to applications running in Azure. Applications can access resources
using this identity instead of or in addition to using keys or passwords.
Security Principals
The first component in RBAC is a security principal. A security principal can be a user, group, service principal, or
managed identity.
Figure 9-2 . Different types of security principals.

User - Any user who has an account in Azure Active Directory is a user.
Group - A collection of users from Azure Active Directory. As a member of a group, a user takes on the roles of
that group in addition to their own.
Service principal - A security identity under which services or applications run.
Managed identity - An Azure Active Directory identity managed by Azure. Managed identities are typically used
when developing cloud applications that manage the credentials for authenticating to Azure services.
The security principal can be applied to most any resource. This means that it's possible to assign a security
principal to a container running within Azure Kubernetes, allowing it to access secrets stored in Key Vault. An Azure
Function could take on a permission allowing it to talk to an Active Directory instance to validate a JWT for a calling
user. Once services are enabled with a service principal, their permissions can be managed granularly using roles
and scopes.
Roles
A security principal can take on many roles or, using a more sartorial analogy, wear many hats. Each role defines a
series of permissions such as "Read messages from Azure Service Bus endpoint". The effective permission set of a
security principal is the combination of all the permissions assigned to all the roles that security principal has.
Azure has a large number of built-in roles and users can define their own roles.
Figure 9-3 . RBAC role definitions.

Built into Azure are also a number of high-level roles such as Owner, Contributor, Reader, and User Account
Administrator. With the Owner role, a security principal can access all resources and assign permissions to others.
A contributor has the same level of access to all resources but they can't assign permissions. A Reader can only
view existing Azure resources and a User Account Administrator can manage access to Azure resources.
More granular built-in roles such as DNS Zone Contributor have rights limited to a single service. Security
principals can take on any number of roles.
Scopes
Roles can be applied to a restricted set of resources within Azure. For instance, applying scope to the previous
example of reading from a Service Bus queue, you can narrow the permission to a single queue: "Read messages
from Azure Service Bus endpoint blah.servicebus.windows.net/queue1 "
The scope can be as narrow as a single resource or it can be applied to an entire resource group, subscription, or
even management group.
When testing if a security principal has a certain permission, the combination of role and scope are taken into
account. This combination provides a powerful authorization mechanism.
Deny
Previously, only "allow" rules were permitted for RBAC. This behavior made some scopes complicated to build. For
instance, allowing a security principal access to all storage accounts except one required granting explicit
permission to a potentially endless list of storage accounts. Every time a new storage account was created, it would
have to be added to this list of accounts. This added management overhead that certainly wasn't desirable.
Deny rules take precedence over allow rules. Now representing the same "allow all but one" scope could be
represented as two rules "allow all" and "deny this one specific one". Deny rules not only ease management but
allow for resources that are extra secure by denying access to everybody.
Checking access
As you can imagine, having a large number of roles and scopes can make figuring out the effective permission of a
service principal quite difficult. Piling deny rules on top of that, only serves to increase the complexity. Fortunately,
there's a permissions calculator that can show the effective permissions for any service principal. It's typically found
under the IAM tab in the portal, as shown in Figure 10-3.
Figure 9-4 . Permission calculator for an app service.

Securing secrets
Passwords and certificates are a common attack vector for attackers. Password-cracking hardware can do a brute-
force attack and try to guess billions of passwords per second. So it's important that the passwords that are used
to access resources are strong, with a large variety of characters. These passwords are exactly the kind of
passwords that are near impossible to remember. Fortunately, the passwords in Azure don't actually need to be
known by any human.
Many security experts suggest that using a password manager to keep your own passwords is the best approach.
While it centralizes your passwords in one location, it also allows using highly complex passwords and ensuring
they're unique for each account. The same system exists within Azure: a central store for secrets.
Azure Key Vault

Azure Key Vault provides a centralized location to store passwords for things such as databases, API keys, and
certificates. Once a secret is entered into the Vault, it's never shown again and the commands to extract and view it
are purposefully complicated. The information in the safe is protected using either software encryption or FIPS
140-2 Level 2 validated Hardware Security Modules.
Access to the key vault is provided through RBACs, meaning that not just any user can access the information in
the vault. Say a web application wishes to access the database connection string stored in Azure Key Vault. To gain
access, applications need to run using a service principal. Under this assumed role, they can read the secrets from
the safe. There are a number of different security settings that can further limit the access that an application has to
the vault, so that it can't update secrets but only read them.
Access to the key vault can be monitored to ensure that only the expected applications are accessing the vault. The
logs can be integrated back into Azure Monitor, unlocking the ability to set up alerts when unexpected conditions
are encountered.
Kubernetes
Within Kubernetes, there's a similar service for maintaining small pieces of secret information. Kubernetes Secrets
can be set via the typical kubectl executable.
Creating a secret is as simple as finding the base64 version of the values to be stored:
echo -n 'admin' | base64

YWRtaW4=
echo -n '1f2d1e2e67df' | base64
MWYyZDFlMmU2N2Rm
Then adding it to a secrets file named secret.yml for example that looks similar to the following example:
apiVersion: v1
kind: Secret
metadata:
name: mysecret
type: Opaque
data:
username: YWRtaW4=
password: MWYyZDFlMmU2N2Rm
Finally, this file can be loaded into Kubernetes by running the following command:
kubectl apply -f ./secret.yaml

These secrets can then be mounted into volumes or exposed to container processes through environment
variables. The Twelve-factor app approach to building applications suggests using the lowest common
denominator to transmit settings to an application. Environment variables are the lowest common denominator,
because they're supported no matter the operating system or application.
An alternative to use the built-in Kubernetes secrets is to access the secrets in Azure Key Vault from within
Kubernetes. The simplest way to do this is to assign an RBAC role to the container looking to load secrets. The
application can then use the Azure Key Vault APIs to access the secrets. However, this approach requires
modifications to the code and doesn't follow the pattern of using environment variables. Instead, it's possible to
inject values into a container through the use of the Azure Key Vault Injector. This approach is actually more secure
than using the Kubernetes secrets directly, as they can be accessed by users on the cluster.
Encryption in transit and at rest

Keeping data safe is important whether it's on disk or transiting between various different services. The most
effective way to keep data from leaking is to encrypt it into a format that can't be easily read by others. Azure
supports a wide range of encryption options.
In transit
There are several ways to encrypt traffic on the network in Azure. The access to Azure services is typically done
over connections that use Transport Layer Security (TLS). For instance, all the connections to the Azure APIs require
TLS connections. Equally, connections to endpoints in Azure storage can be restricted to work only over TLS
encrypted connections.
TLS is a complicated protocol and simply knowing that the connection is using TLS isn't sufficient to ensure
security. For instance, TLS 1.0 is chronically insecure, and TLS 1.1 isn't much better. Even within the versions of TLS,
there are various settings that can make the connections easier to decrypt. The best course of action is to check and
see if the server connection is using up-to-date and well configured protocols.
This check can be done by an external service such as SSL labs' SSL Server Test. A test run against a typical Azure
endpoint, in this case a service bus endpoint, yields a near perfect score of A.
Even services like Azure SQL databases use TLS encryption to keep data hidden. The interesting part about
encrypting the data in transit using TLS is that it isn't possible, even for Microsoft, to listen in on the connection
between computers running TLS. This should provide comfort for companies concerned that their data may be at
risk from Microsoft proper or even a state actor with more resources than the standard attacker.
Figure 9-5 . SSL labs report showing a score of A for a Service Bus endpoint.
While this level of encryption isn't going to be sufficient for all time, it should inspire confidence that Azure TLS
connections are quite secure. Azure will continue to evolve its security standards as encryption improves. It's nice
to know that there's somebody watching the security standards and updating Azure as they improve.
At rest
In any application, there are a number of places where data rests on disk. The application code itself is loaded from
some storage mechanism. Most applications also use some kind of database such as SQL Server, Cosmos DB, or
even the amazingly price-efficient Table Storage. These databases all use heavily encrypted storage to ensure that
nobody other than the applications with proper permissions can read your data. Even the system operators can't
read data that has been encrypted. So customers can remain confident their secret information remains secret.
Storage
The underpinning of much of Azure is the Azure Storage engine. Virtual machine disks are mounted on top of
Azure Storage. Azure Kubernetes Services run on virtual machines that, themselves, are hosted on Azure Storage.
Even serverless technologies, such as Azure Functions Apps and Azure Container Instances, run out of disk that is
part of Azure Storage.
If Azure Storage is well encrypted, then it provides for a foundation for most everything else to also be encrypted.
Azure Storage is encrypted with FIPS 140-2 compliant 256-bit AES. This is a well-regarded encryption technology
having been the subject of extensive academic scrutiny over the last 20 or so years. At present, there's no known
practical attack that would allow someone without knowledge of the key to read data encrypted by AES.
By default, the keys used for encrypting Azure Storage are managed by Microsoft. There are extensive protections
in place to ensure to prevent malicious access to these keys. However, users with particular encryption
requirements can also provide their own storage keys that are managed in Azure Key Vault. These keys can be
revoked at any time, which would effectively render the contents of the Storage account using them inaccessible.
Virtual machines use encrypted storage, but it's possible to provide another layer of encryption by using
technologies like BitLocker on Windows or DM-Crypt on Linux. These technologies mean that even if the disk
image was leaked off of storage, it would remain near impossible to read it.
Azure SQL
Databases hosted on Azure SQL use a technology called Transparent Data Encryption (TDE) to ensure data remains
encrypted. It's enabled by default on all newly created SQL databases, but must be enabled manually for legacy
databases. TDE executes real-time encryption and decryption of not just the database, but also the backups and
transaction logs.
The encryption parameters are stored in the master database and, on startup, are read into memory for the
remaining operations. This means that the master database must remain unencrypted. The actual key is managed
by Microsoft. However, users with exacting security requirements may provide their own key in Key Vault in much
the same way as is done for Azure Storage. The Key Vault provides for such services as key rotation and revocation.
The "Transparent" part of TDS comes from the fact that there aren't client changes needed to use an encrypted
database. While this approach provides for good security, leaking the database password is enough for users to be
able to decrypt the data. There's another approach that encrypts individual columns or tables in a database. Always
Encrypted ensures that at no point the encrypted data appears in plain text inside the database.
Setting up this tier of encryption requires running through a wizard in SQL Server Management Studio to select
the sort of encryption and where in Key Vault to store the associated keys.
Figure 9-6 . Selecting columns in a table to be encrypted using Always Encrypted.
Client applications that read information from these encrypted columns need to make special allowances to read
encrypted data. Connection strings need to be updated with Column Encryption Setting=Enabled and client
credentials must be retrieved from the Key Vault. The SQL Server client must then be primed with the column
encryption keys. Once that is done, the remaining actions use the standard interfaces to SQL Client. That is, tools
like Dapper and Entity Framework, which are built on top of SQL Client, will continue to work without changes.
Always Encrypted may not yet be available for every SQL Server driver on every language.
The combination of TDE and Always Encrypted, both of which can be used with client-specific keys, ensures that
even the most exacting encryption requirements are supported.
Cosmos DB
Cosmos DB is the newest database provided by Microsoft in Azure. It has been built from the ground up with
security and cryptography in mind. AES-256bit encryption is standard for all Cosmos DB databases and can't be
disabled. Coupled with the TLS 1.2 requirement for communication, the entire storage solution is encrypted.
Figure 9-7 . The flow of data encryption within Cosmos DB.
While Cosmos DB doesn't provide for supplying customer encryption keys, there has been significant work done
by the team to ensure it remains PCI-DSS compliant without that. Cosmos DB also doesn't support any sort of
single column encryption similar to Azure SQL's Always Encrypted yet.
Keeping secure
Azure has all the tools necessary to release a highly secure product. However, a chain is only as strong as its
weakest link. If the applications deployed on top of Azure aren't developed with a proper security mindset and
good security audits, then they become the weak link in the chain. There are many great static analysis tools,
encryption libraries, and security practices that can be used to ensure that the software installed on Azure is as
secure as Azure itself. Examples include static analysis tools, encryption libraries, and security practices.
PR EVIO U S NE XT
DevOps
The favorite mantra of software consultants is to answer "It depends" to any question posed. It isn't because
software consultants are fond of not taking a position. It's because there's no one true answer to any questions in
software. There's no absolute right and wrong, but rather a balance between opposites.
Take, for instance, the two major schools of developing web applications: Single Page Applications (SPAs) versus
server-side applications. On the one hand, the user experience tends to be better with SPAs and the amount of
traffic to the web server can be minimized making it possible to host them on something as simple as static
hosting. On the other hand, SPAs tend to be slower to develop and more difficult to test. Which one is the right
choice? Well, it depends on your situation.
Cloud-native applications aren't immune to that same dichotomy. They have clear advantages in terms of speed of
development, stability, and scalability, but managing them can be quite a bit more difficult.
Years ago, it wasn't uncommon for the process of moving an application from development to production to take a
month, or even more. Companies released software on a 6-month or even every year cadence. One needs to look
no further than Microsoft Windows to get an idea for the cadence of releases that were acceptable before the ever-
green days of Windows 10. Five years passed between Windows XP and Vista, a further 3 between Vista and
Windows 7.
It's now fairly well established that being able to release software rapidly gives fast-moving companies a huge
market advantage over their more sloth-like competitors. It's for that reason that major updates to Windows 10
are now approximately every six months.
The patterns and practices that enable faster, more reliable releases to deliver value to the business are collectively
known as DevOps. They consist of a wide range of ideas spanning the entire software development life cycle from
specifying an application all the way up to delivering and operating that application.
DevOps emerged before microservices and it's likely that the movement towards smaller, more fit to purpose
services wouldn't have been possible without DevOps to make releasing and operating not just one but many
applications in production easier.
Figure 10-1 - DevOps and microservices.
Through good DevOps practices, it's possible to realize the advantages of cloud-native applications without
suffocating under a mountain of work actually operating the applications.
There's no golden hammer when it comes to DevOps. Nobody can sell a complete and all-encompassing solution
for releasing and operating high-quality applications. This is because each application is wildly different from all
others. However, there are tools that can make DevOps a far less daunting proposition. One of these tools is known
as Azure DevOps.
Azure DevOps
Azure DevOps has a long pedigree. It can trace its roots back to when Team Foundation Server first moved online
and through the various name changes: Visual Studio Online and Visual Studio Team Services. Through the years,
however, it has become far more than its predecessors.
Azure DevOps is divided into five major components:
Figure 10-2 - Azure DevOps.

Azure Repos - Source code management that supports the venerable Team Foundation Version Control (TFVC)
and the industry favorite Git. Pull requests provide a way to enable social coding by fostering discussion of
changes as they're made.
Azure Boards - Provides an issue and work item tracking tool that strives to allow users to pick the workflows
that work best for them. It comes with a number of pre-configured templates including ones to support SCRUM
and Kanban styles of development.
Azure Pipelines - A build and release management system that supports tight integration with Azure. Builds can
be run on a variety of platforms from Windows to Linux to MacOS. Build agents may be provisioned in the cloud
or on-premises.
Azure Test Plans - No QA person will be left behind with the test management and exploratory testing support
offered by the Test Plans feature.
Azure Ar tifacts - An artifact feed that allows companies to create their own, internal, versions of NuGet, npm,
and others. It serves a double purpose of acting as a cache of upstream packages if there's a failure of a centralized
repository.
The top-level organizational unit in Azure DevOps is known as a Project. Within each project the various
components, such as Azure Artifacts, can be turned on and off. Each of these components provides different
advantages for cloud-native applications. The three most useful are repositories, boards, and pipelines. If users
want to manage their source code in another repository stack, such as GitHub, but still take advantage of Azure
Pipelines and other components, that's perfectly possible.
Fortunately, development teams have many options when selecting a repository. One of them is GitHub.
GitHub Actions
Founded in 2009, GitHub is a widely popular web-based repository for hosting projects, documentation, and code.
Many large tech companies, such as Apple, Amazon, Google, and mainstream corporations use GitHub. GitHub
uses the open-source, distributed version control system named Git as its foundation. On top, it then adds its own
set of features, including defect tracking, feature and pull requests, tasks management, and wikis for each code
base.
As GitHub evolves, it too is adding DevOps features. For example, GitHub has its own continuous
integration/continuous delivery (CI/CD) pipeline, called GitHub Actions . GitHub Actions is a community-powered
workflow automation tool. It lets DevOps teams integrate with their existing tooling, mix and match new products,
and hook into their software lifecycle, including existing CI/CD partners."
GitHub has over 40 million users, making it the largest host of source code in the world. In October of 2018,
Microsoft purchased GitHub. Microsoft has pledged that GitHub will remain an open platform that any developer
can plug into and extend. It continues to operate as an independent company. GitHub offers plans for enterprise,
team, professional, and free accounts.
Source control
Organizing the code for a cloud-native application can be challenging. Instead of a single giant application, the
cloud-native applications tend to be made up of a web of smaller applications that talk with one another. As with
all things in computing, the best arrangement of code remains an open question. There are examples of successful
applications using different kinds of layouts, but two variants seem to have the most popularity.
Before getting down into the actual source control itself, it's probably worth deciding on how many projects are
appropriate. Within a single project, there's support for multiple repositories, and build pipelines. Boards are a little
more complicated, but there too, the tasks can easily be assigned to multiple teams within a single project. It's
possible to support hundreds, even thousands of developers, out of a single Azure DevOps project. Doing so is
likely the best approach as it provides a single place for all developer to work out of and reduces the confusion of
finding that one application when developers are unsure in which project in which it resides.
Splitting up code for microservices within the Azure DevOps project can be slightly more challenging.
Figure 10-3 - One vs. many repositories.

Repository per microservice
At first glance, this seems like the most logical approach to splitting up the source code for microservices. Each
repository can contain the code needed to build the one microservice. The advantages to this approach are readily
visible:
1. Instructions for building and maintaining the application can be added to a README file at the root of each
repository. When flipping through the repositories, it's easy to find these instructions, reducing spin-up time for
developers.
2. Every service is located in a logical place, easily found by knowing the name of the service.
3. Builds can easily be set up such that they're only triggered when a change is made to the owning repository.
4. The number of changes coming into a repository is limited to the small number of developers working on the
project.
5. Security is easy to set up by restricting the repositories to which developers have read and write permissions.
6. Repository level settings can be changed by the owning team with a minimum of discussion with others.
One of the key ideas behind microservices is that services should be siloed and separated from each other. When
using Domain Driven Design to decide on the boundaries for services the services act as transactional boundaries.
Database updates shouldn't span multiple services. This collection of related data is referred to as a bounded
context. This idea is reflected by the isolation of microservice data to a database separate and autonomous from
the rest of the services. It makes a great deal of sense to carry this idea all the way through to the source code.
However, this approach isn't without its issues. One of the more gnarly development problems of our time is
managing dependencies. Consider the number of files that make up the average node_modules directory. A fresh
install of something like create-react-app is likely to bring with it thousands of packages. The question of how to
manage these dependencies is a difficult one.
If a dependency is updated, then downstream packages must also update this dependency. Unfortunately, that
takes development work so, invariably, the node_modules directory ends up with multiple versions of a single
package, each one a dependency of some other package that is versioned at a slightly different cadence. When
deploying an application, which version of a dependency should be used? The version that is currently in
production? The version that is currently in Beta but is likely to be in production by the time the consumer makes it
to production? Difficult problems that aren't resolved by just using microservices.
There are libraries that are depended upon by a wide variety of projects. By dividing the microservices up with one
in each repository the internal dependencies can best be resolved by using the internal repository, Azure Artifacts.
Builds for libraries will push their latest versions into Azure Artifacts for internal consumption. The downstream
project must still be manually updated to take a dependency on the newly updated packages.
Another disadvantage presents itself when moving code between services. Although it would be nice to believe
that the first division of an application into microservices is 100% correct, the reality is that rarely we're so
prescient as to make no service division mistakes. Thus, functionality and the code that drives it will need to move
from service to service: repository to repository. When leaping from one repository to another, the code loses its
history. There are many cases, especially in the event of an audit, where having full history on a piece of code is
invaluable.
The final and most important disadvantage is coordinating changes. In a true microservices application, there
should be no deployment dependencies between services. It should be possible to deploy services A, B, and C in
any order as they have loose coupling. In reality, however, there are times when it's desirable to make a change
that crosses multiple repositories at the same time. Some examples include updating a library to close a security
hole or changing a communication protocol used by all services.
To do a cross-repository change requires a commit to each repository be made in succession. Each change in each
repository will need to be pull-requested and reviewed separately. This can be difficult to coordinate.
An alternative to using many repositories is to put all the source code together in a giant, all knowing, single
repository.
Single repository
In this approach, sometimes referred to as a monorepository, all the source code for every service is put into the
same repository. At first, this seems like a terrible idea likely to make dealing with source code unwieldy. There are,
however, some marked advantages to working this way.
The first advantage is that it's easier to manage dependencies between projects. Instead of relying on some
external artifact feed, projects can directly import one another. This means that updates are instant, and conflicting
versions are likely to be found at compile time on the developer's workstation. In effect, shifting some of the
integration testing left.
When moving code between projects, it's now easier to preserve the history as the files will be detected as having
been moved rather than being rewritten.
Another advantage is that wide ranging changes that cross service boundaries can be made in a single commit.
This reduces the overhead of having potentially dozens of changes to review individually.
There are many tools that can perform static analysis of code to detect insecure programming practices or
problematic use of APIs. In a multi-repository world, each repository will need to be iterated over to find the
problems in them. The single repository allows running the analysis all in one place.
There are also many disadvantages to the single repository approach. One of the most worrying ones is that
having a single repository raises security concerns. If the contents of a repository are leaked in a repository per
service model, the amount of code lost is minimal. With a single repository, everything the company owns could
be lost. There have been many examples in the past of this happening and derailing entire game development
efforts. Having multiple repositories exposes less surface area, which is a desirable trait in most security practices.
The size of the single repository is likely to become unmanageable rapidly. This presents some interesting
performance implications. It may become necessary to use specialized tools such as Virtual File System for Git,
which was originally designed to improve the experience for developers on the Windows team.
Frequently the argument for using a single repository boils down to an argument that Facebook or Google use this
method for source code arrangement. If the approach is good enough for these companies, then, surely, it's the
correct approach for all companies. The truth of the matter is that few companies operate on anything like the
scale of Facebook or Google. The problems that occur at those scales are different from those most developers will
face. What is good for the goose may not be good for the gander.
In the end, either solution can be used to host the source code for microservices. However, in most cases, the
management, and engineering overhead of operating in a single repository isn't worth the meager advantages.
Splitting code up over multiple repositories encourages better separation of concerns and encourages autonomy
among development teams.
Standard directory structure
Regardless of the single versus multiple repositories debate each service will have its own directory. One of the
best optimizations to allow developers to cross between projects quickly is to maintain a standard directory
structure.
Figure 10-4 - Standard directory structure.
Whenever a new project is created, a template that puts in place the correct structure should be used. This
template can also include such useful items as a skeleton README file and an azure-pipelines.yml . In any
microservice architecture, a high degree of variance between projects makes bulk operations against the services
more difficult.
There are many tools that can provide templating for an entire directory, containing several source code
directories. Yeoman is popular in the JavaScript world and GitHub have recently released Repository Templates,
which provide much of the same functionality.
Task management
Managing tasks in any project can be difficult. Up front there are countless questions to be answered about the
sort of workflows to set up to ensure optimal developer productivity.
Cloud-native applications tend to be smaller than traditional software products or at least they're divided into
smaller services. Tracking of issues or tasks related to these services remains as important as with any other
software project. Nobody wants to lose track of some work item or explain to a customer that their issue wasn't
properly logged. Boards are configured at the project level but within each project, areas can be defined. These
allow breaking down issues across several components. The advantage to keeping all the work for the entire
application in one place is that it's easy to move work items from one team to another as they're understood better.
Azure DevOps comes with a number of popular templates pre-configured. In the most basic configuration, all that
is needed to know is what's in the backlog, what people are working on, and what's done. It's important to have
this visibility into the process of building software, so that work can be prioritized and completed tasks reported to
the customer. Of course, few software projects stick to a process as simple as to do , doing , and done . It doesn't
take long for people to start adding steps like QA or Detailed Specification to the process.
One of the more important parts of Agile methodologies is self-introspection at regular intervals. These reviews
are meant to provide insight into what problems the team is facing and how they can be improved. Frequently, this
means changing the flow of issues and features through the development process. So, it's perfectly healthy to
expand the layouts of the boards with additional stages.
The stages in the boards aren't the only organizational tool. Depending on the configuration of the board, there's a
hierarchy of work items. The most granular item that can appear on a board is a task. Out of the box a task
contains fields for a title, description, a priority, an estimate of the amount of work remaining and the ability to link
to other work items or development items (branches, commits, pull requests, builds, and so forth). Work items can
be classified into different areas of the application and different iterations (sprints) to make finding them easier.
Figure 10-5 - Task in Azure DevOps.
The description field supports the normal styles you'd expect (bold, italic underscore and strike through) and the
ability to insert images. This makes it a powerful tool for use when specifying work or bugs.
Tasks can be rolled up into features, which define a larger unit of work. Features, in turn, can be rolled up into
epics. Classifying tasks in this hierarchy makes it much easier to understand how close a large feature is to rolling
out.
Figure 10-6 - Work item in Azure DevOps.

There are different kinds of views into the issues in Azure Boards. Items that aren't yet scheduled appear in the
backlog. From there, they can be assigned to a sprint. A sprint is a time box during which it's expected some
quantity of work will be completed. This work can include tasks but also the resolution of tickets. Once there, the
entire sprint can be managed from the Sprint board section. This view shows how work is progressing and
includes a burn down chart to give an ever-updating estimate of if the sprint will be successful.
Figure 10-7 - Board in Azure DevOps.

By now, it should be apparent that there's a great deal of power in the Boards in Azure DevOps. For developers,
there are easy views of what is being worked on. For project managers views into upcoming work as well as an
overview of existing work. For managers, there are plenty of reports about resourcing and capacity. Unfortunately,
there's nothing magical about cloud-native applications that eliminate the need to track work. But if you must track
work, there are a few places where the experience is better than in Azure DevOps.
CI/CD pipelines
Almost no change in the software development life cycle has been so revolutionary as the advent of continuous
integration (CI) and continuous delivery (CD). Building and running automated tests against the source code of a
project as soon as a change is checked in catches mistakes early. Prior to the advent of continuous integration
builds, it wouldn't be uncommon to pull code from the repository and find that it didn't pass tests or couldn't even
be built. This resulted in tracking down the source of the breakage.
Traditionally shipping software to the production environment required extensive documentation and a list of
steps. Each one of these steps needed to be manually completed in a very error prone process.
Figure 10-8 - Checklist.

The sister of continuous integration is continuous delivery in which the freshly built packages are deployed to an
environment. The manual process can't scale to match the speed of development so automation becomes more
important. Checklists are replaced by scripts that can execute the same tasks faster and more accurately than any
human.
The environment to which continuous delivery delivers might be a test environment or, as is being done by many
major technology companies, it could be the production environment. The latter requires an investment in high-
quality tests that can give confidence that a change isn't going to break production for users. In the same way that
continuous integration caught issues in the code early continuous delivery catches issues in the deployment
process early.
The importance of automating the build and delivery process is accentuated by cloud-native applications.
Deployments happen more frequently and to more environments so manually deploying borders on impossible.
Azure Builds
Azure DevOps provides a set of tools to make continuous integration and deployment easier than ever. These tools
are located under Azure Pipelines. The first of them is Azure Builds, which is a tool for running YAML-based build
definitions at scale. Users can either bring their own build machines (great for if the build requires a meticulously
set up environment) or use a machine from a constantly refreshed pool of Azure hosted virtual machines. These
hosted build agents come pre-installed with a wide range of development tools for not just .NET development but
for everything from Java to Python to iPhone development.
DevOps includes a wide range of out of the box build definitions that can be customized for any build. The build
definitions are defined in a file called azure-pipelines.yml and checked into the repository so they can be
versioned along with the source code. This makes it much easier to make changes to the build pipeline in a branch
as the changes can be checked into just that branch. An example azure-pipelines.yml for building an ASP.NET web
application on full framework is show in Figure 10-9.
name: $(rev:r)
variables:
version: 9.2.0.$(Build.BuildNumber)
solution: Portals.sln
artifactName: drop
buildPlatform: any cpu
buildConfiguration: release
pool:
name: Hosted VS2017
demands:
- msbuild
- visualstudio
- vstest
steps:
- task: NuGetToolInstaller@0
displayName: 'Use NuGet 4.4.1'
inputs:
versionSpec: 4.4.1
- task: NuGetCommand@2
displayName: 'NuGet restore'
inputs:
restoreSolution: '$(solution)'
- task: VSBuild@1
displayName: 'Build solution'
inputs:
solution: '$(solution)'
msbuildArgs: '-p:DeployOnBuild=true -p:WebPublishMethod=Package -p:PackageAsSingleFile=true -
p:SkipInvalidConfigurations=true -p:PackageLocation="$(build.artifactstagingdirectory)\\"'
platform: '$(buildPlatform)'
configuration: '$(buildConfiguration)'
- task: VSTest@2
displayName: 'Test Assemblies'
inputs:
testAssemblyVer2: |
**\$(buildConfiguration)\**\*test*.dll
!**\obj\**
!**\*testadapter.dll
platform: '$(buildPlatform)'
configuration: '$(buildConfiguration)'
- task: CopyFiles@2
displayName: 'Copy UI Test Files to: $(build.artifactstagingdirectory)'
inputs:
SourceFolder: UITests
TargetFolder: '$(build.artifactstagingdirectory)/uitests'
- task: PublishBuildArtifacts@1
displayName: 'Publish Artifact'
inputs:
PathtoPublish: '$(build.artifactstagingdirectory)'
ArtifactName: '$(artifactName)'
condition: succeededOrFailed()
Figure 10-9 - A sample azure-pipelines.yml

This build definition uses a number of built-in tasks that make creating builds as simple as building a Lego set
(simpler than the giant Millennium Falcon). For instance, the NuGet task restores NuGet packages, while the
VSBuild task calls the Visual Studio build tools to perform the actual compilation. There are hundreds of different
tasks available in Azure DevOps, with thousands more that are maintained by the community. It's likely that no
matter what build tasks you're looking to run, somebody has built one already.
Builds can be triggered manually, by a check-in, on a schedule, or by the completion of another build. In most
cases, building on every check-in is desirable. Builds can be filtered so that different builds run against different
parts of the repository or against different branches. This allows for scenarios like running fast builds with reduced
testing on pull requests and running a full regression suite against the trunk on a nightly basis.
The end result of a build is a collection of files known as build artifacts. These artifacts can be passed along to the
next step in the build process or added to an Azure Artifact feed, so they can be consumed by other builds.
Azure DevOps releases
Builds take care of compiling the software into a shippable package, but the artifacts still need to be pushed out to
a testing environment to complete continuous delivery. For this, Azure DevOps uses a separate tool called
Releases. The Releases tool makes use of the same tasks' library that were available to the Build but introduce a
concept of "stages". A stage is an isolated environment into which the package is installed. For instance, a product
might make use of a development, a QA, and a production environment. Code is continuously delivered into the
development environment where automated tests can be run against it. Once those tests pass the release moves
onto the QA environment for manual testing. Finally, the code is pushed to production where it's visible to
everybody.
Figure 10-10 - Release pipeline

Each stage in the build can be automatically triggered by the completion of the previous phase. In many cases,
however, this isn't desirable. Moving code into production might require approval from somebody. The Releases
tool supports this by allowing approvers at each step of the release pipeline. Rules can be set up such that a
specific person or group of people must sign off on a release before it makes into production. These gates allow
for manual quality checks and also for compliance with any regulatory requirements related to control what goes
into production.
Everybody gets a build pipeline
There's no cost to configuring many build pipelines, so it's advantageous to have at least one build pipeline per
microservice. Ideally, microservices are independently deployable to any environment so having each one able to
be released via its own pipeline without releasing a mass of unrelated code is perfect. Each pipeline can have its
own set of approvals allowing for variations in build process for each service.
Versioning releases
One drawback to using the Releases functionality is that it can't be defined in a checked-in azure-pipelines.yml
file. There are many reasons you might want to do that from having per-branch release definitions to including a
release skeleton in your project template. Fortunately, work is ongoing to shift some of the stages support into the
Build component. This will be known as multi-stage build and the first version is available now!
PR EVIO U S NE XT
Feature flags
In chapter 1, we affirmed that cloud native is much about speed and agility. Users expect rapid responsiveness,
innovative features, and zero downtime. Feature flags are a modern deployment technique that helps increase
agility for cloud-native applications. They enable you to deploy new features into a production environment, but
restrict their availability. With the flick of a switch, you can activate a new feature for specific users without
restarting the app or deploying new code. They separate the release of new features from their code deployment.
Feature flags are built upon conditional logic that control visibility of functionality for users at runtime. In modern
cloud-native systems, it's common to deploy new features into production early, but test them with a limited
audience. As confidence increases, the feature can be incrementally rolled out to wider audiences.
Other use cases for feature flags include:
Restrict premium functionality to specific customer groups willing to pay higher subscription fees.
Stabilize a system by quickly deactivating a problem feature, avoiding the risks of a rollback or immediate
hotfix.
Disable an optional feature with high resource consumption during peak usage periods.
Conduct experimental feature releases to small user segments to validate feasibility and popularity.
Feature flags also promote trunk-based development. It's a source-control branching model where developers
collaborate on features in a single branch. The approach minimizes the risk and complexity of merging large
numbers of long-running feature branches. Features are unavailable until activated.
Implementing feature flags

At its core, a feature flag is a reference to a simple decision object . It returns a Boolean state of on or off . The
flag typically wraps a block of code that encapsulates a feature capability. The state of the flag determines whether
that code block executes for a given user. Figure 10-11 shows the implementation.
if (featureFlag) {
// Run this code block if the featureFlag value is true
} else {
// Run this code block if the featureFlag value is false
}
Figure 10-11 - Simple feature flag implementation.

Note how this approach separates the decision logic from the feature code.
In chapter 1, we discussed the Twelve-Factor App . The guidance recommended keeping configuration settings
external from application executable code. When needed, settings can be read in from the external source. Feature
flag configuration values should also be independent from their codebase. By externalizing flag configuration in a
separate repository, you can change flag state without modifying and redeploying the application.
Azure App Configuration provides a centralized repository for feature flags. With it, you define different kinds of
feature flags and manipulate their states quickly and confidently. You add the App Configuration client libraries to
your application to enable feature flag functionality. Various programming language frameworks are supported.
Feature flags can be easily implemented in an ASP.NET Core service. Installing the .NET Feature Management
libraries and App Configuration provider enable you to declaratively add feature flags to your code. They enable
FeatureGate attributes so that you don't have to manually write if statements across your codebase.
Once configured in your Startup class, you can add feature flag functionality at the controller, action, or middleware
level. Figure 10-12 presents controller and action implementation:
[FeatureGate(MyFeatureFlags.FeatureA)]
public class ProductController : Controller
{
...
}
[FeatureGate(MyFeatureFlags.FeatureA)]
public IActionResult UpdateProductStatus()
{
return ObjectResult(ProductDto);
}
Figure 10-12 - Feature flag implementation in a controller and action.

If a feature flag is disabled, the user will receive a 404 (Not Found) status code with no response body.
Feature flags can also be injected directly into C# classes. Figure 10-13 shows feature flag injection:
public class ProductController : Controller

{
private readonly IFeatureManager _featureManager;
public ProductController(IFeatureManager featureManager)

{
_featureManager = featureManager;
}
}
Figure 10-13 - Feature flag injection into a class.

The Feature Management libraries manage the feature flag lifecycle behind the scenes. For example, to minimize
high numbers of calls to the configuration store, the libraries cache flag states for a specified duration. They can
guarantee the immutability of flag states during a request call. They also offer a Point-in-time snapshot . You can
reconstruct the history of any key-value and provide its past value at any moment within the previous seven days.
PR EVIO U S NE XT
Infrastructure as code
Cloud-native systems embrace microservices, containers, and modern system design to achieve speed and agility.
They provide automated build and release stages to ensure consistent and quality code. But, that's only part of the
story. How do you provision the cloud environments upon which these systems run?
Modern cloud-native applications embrace the widely accepted practice of Infrastructure as Code, or IaC . With
IaC, you automate platform provisioning. You essentially apply software engineering practices such as testing and
versioning to your DevOps practices. Your infrastructure and deployments are automated, consistent, and
repeatable. Just as continuous delivery automated the traditional model of manual deployments, Infrastructure as
Code (IaC) is evolving how application environments are managed.
Tools like Azure Resource Manager (ARM), Terraform, and the Azure Command Line Interface (CLI) enable you to
declaratively script the cloud infrastructure you require.
Azure Resource Manager templates

ARM stands for Azure Resource Manager. It's an API provisioning engine that is built into Azure and exposed as an
API service. ARM enables you to deploy, update, delete, and manage the resources contained in Azure resource
group in a single, coordinated operation. You provide the engine with a JSON-based template that specifies the
resources you require and their configuration. ARM automatically orchestrates the deployment in the correct order
respecting dependencies. The engine ensures idempotency. If a desired resource already exists with the same
configuration, provisioning will be ignored.
Azure Resource Manager templates are a JSON-based language for defining various resources in Azure. The basic
schema looks something like Figure 10-14.
{
"$schema": "https://schema.management.azure.com/schemas/2015-01-01/deploymentTemplate.json#",
"contentVersion": "",
"apiProfile": "",
"parameters": { },
"variables": { },
"functions": [ ],
"resources": [ ],
"outputs": { }
}
Figure 10-14 - The schema for a Resource Manager template

Within this template, one might define a storage container inside the resources section like so:
"resources": [
{
"type": "Microsoft.Storage/storageAccounts",
"name": "[variables('storageAccountName')]",
"location": "[parameters('location')]",
"apiVersion": "2018-07-01",
"sku": {
"name": "[parameters('storageAccountType')]"
},
"kind": "StorageV2",
"properties": {}
}
],
Figure 10-15 - An example of a storage account defined in a Resource Manager template

An ARM template can be parameterized with dynamic environment and configuration information. Doing so
enables it to be reused to define different environments, such as development, QA, or production. Normally, the
template creates all resources within a single Azure resource group. It's possible to define multiple resource
groups in a single Resource Manager template, if needed. You can delete all resources in an environment by
deleting the resource group itself. Cost analysis can also be run at the resource group level, allowing for quick
accounting of how much each environment is costing.
There are many examples or ARM templates available in the Azure Quickstart Templates project on GitHub. They
can help accelerate creating a new template or modifying an existing one.
Resource Manager templates can be run in many of ways. Perhaps the simplest way is to simply paste them into
the Azure portal. For experimental deployments, this method can be quick. They can also be run as part of a build
or release process in Azure DevOps. There are tasks that will leverage connections into Azure to run the templates.
Changes to Resource Manager templates are applied incrementally, meaning that to add a new resource requires
just adding it to the template. The tooling will reconcile differences between the current resources and those
defined in the template. Resources will then be created or altered so they match what is defined in the template.
Terraform
Cloud-native applications are often constructed to be cloud agnostic . Being so means the application isn't tightly
coupled to a particular cloud vendor and can be deployed to any public cloud.
Terraform is commercial templating tool that can provision cloud-native applications across all the major cloud
players: Azure, Google Cloud Platform, AWS, and AliCloud. Instead of using JSON as the template definition
language, it uses the slightly more terse YAML.
An example Terraform file that does the same as the previous Resource Manager template (Figure 10-15) is shown
in Figure 10-16:
provider "azurerm" {
version = "=1.28.0"
}
resource "azurerm_resource_group" "test" {

name = "production"
location = "West US"
}
resource "azurerm_storage_account" "testsa" {

name = "${var.storageAccountName}"
resource_group_name = "${azurerm_resource_group.testrg.name}"
location = "${var.region}"
account_tier = "${var.tier}"
account_replication_type = "${var.replicationType}"
Figure 10-16 - An example of a Resource Manager template

Terraform also provides intuitive error messages for problem templates. There's also a handy validate task that can
be used in the build phase to catch template errors early.
As with Resource Manager templates, command-line tools are available to deploy Terraform templates. There are
also community-created tasks in Azure Pipelines that can validate and apply Terraform templates.
Sometimes Terraform and ARM templates output meaningful values, such as a connection string to a newly
created database. This information can be captured in the build pipeline and used in subsequent tasks.
Azure CLI Scripts and Tasks

Finally, you can leverage Azure CLI to declaratively script your cloud infrastructure. Azure CLI scripts can be
created, found, and shared to provision and configure almost any Azure resource. The CLI is simple to use with a
gentle learning curve. Scripts are executed within either PowerShell or Bash. They're also straightforward to debug,
especially when compared with ARM templates.
Azure CLI scripts work well when you need to tear down and redeploy your infrastructure. Updating an existing
environment can be tricky. Many CLI commands aren't idempotent. That means they'll recreate the resource each
time they're run, even if the resource already exists. It's always possible to add code that checks for the existence of
each resource before creating it. But, doing so, your script can become bloated and difficult to manage.
These scripts can also be embedded in Azure DevOps pipelines as Azure CLI tasks . Executing the pipeline invokes
the script.
Figure 10-17 shows a YAML snippet that lists the version of Azure CLI and the details of the subscription. Note how
Azure CLI commands are included in an inline script.
- task: AzureCLI@2
displayName: Azure CLI
inputs:
azureSubscription: <Name of the Azure Resource Manager service connection>
scriptType: ps
scriptLocation: inlineScript
inlineScript: |
az --version
az account show
Figure 10-17 - Azure CLI script

In the article, What is Infrastructure as Code, Author Sam Guckenheimer describes how, "Teams who implement
IaC can deliver stable environments rapidly and at scale. Teams avoid manual configuration of environments and
enforce consistency by representing the desired state of their environments via code. Infrastructure deployments
with IaC are repeatable and prevent runtime issues caused by configuration drift or missing dependencies. DevOps
teams can work together with a unified set of practices and tools to deliver applications and their supporting
infrastructure rapidly, reliably, and at scale."
PR EVIO U S NE XT
Cloud Native Application Bundles
A key property of cloud-native applications is that they leverage the capabilities of the cloud to speed up
development. This design often means that a full application uses different kinds of technologies. Applications may
be shipped in Docker containers, some services may use Azure Functions, while other parts may run directly on
virtual machines allocated on large metal servers with hardware GPU acceleration. No two cloud-native
applications are the same, so it's been difficult to provide a single mechanism for shipping them.
The Docker containers may run on Kubernetes using a Helm Chart for deployment. The Azure Functions may be
allocated using Terraform templates. Finally, the virtual machines may be allocated using Terraform but built out
using Ansible. This is a whole mess of technologies and there has been no way to package them all together into a
reasonable package. Until now.
Cloud Native Application Bundles (CNABs) are a joint effort by a number of community-minded companies such
as Microsoft, Docker, and HashiCorp to develop a specification to package distributed applications.
The effort was announced in December of 2018, so there's still a fair bit of work to do to expose the effort to the
greater community. However, there's already an open specification and a reference implementation known as
Duffle. This tool, which was written in Go, is a joint effort between Docker and Microsoft.
The CNABs can contain different kinds of installation technologies. This allows things like Helm Charts, Terraform
templates, and Ansible Playbooks to coexist in the same package. Once built, the packages are self-contained and
portable; they can be installed from a USB stick. The packages are cryptographically signed to ensure they
originate from the party they claim.
The core of a CNAB is a file called bundle.json . This file defines the contents of the bundle, be they Terraform or
images or anything else. Figure 11-9 defines a CNAB that invokes some Terraform. Notice, however, that it actually
defines an invocation image that is used to invoke the Terraform. When packaged up, the Docker file that is located
in the cnab directory is built into a Docker image, which will be included in the bundle. Having Terraform installed
inside a Docker container in the bundle means that users don't need to have Terraform installed on their machine
to run the bundling.
{
"name": "terraform",
"version": "0.1.0",
"schemaVersion": "v1.0.0-WD",
"parameters": {
"backend": {
"type": "boolean",
"defaultValue": false,
"destination": {
"env": "TF_VAR_backend"
}
}
},
"invocationImages": [
{
"imageType": "docker",
"image": "cnab/terraform:latest"
}
],
"credentials": {
"tenant_id": {
"env": "TF_VAR_tenant_id"
},
"client_id": {
"env": "TF_VAR_client_id"
},
"client_secret": {
"env": "TF_VAR_client_secret"
},
"subscription_id": {
"env": "TF_VAR_subscription_id"
},
"ssh_authorized_key": {
"env": "TF_VAR_ssh_authorized_key"
}
},
"actions": {
"status": {
"modifies": true
}
}
}
Figure 10-18 - An example Terraform file

The bundle.json also defines a set of parameters that are passed down into the Terraform. Parameterization of the
bundle allows for installation in a variety of different environments.
The CNAB format is also flexible, allowing it to be used against any cloud. It can even be used against on-premises
solutions such as OpenStack.
DevOps Decisions
There are so many great tools in the DevOps space these days and even more fantastic books and papers on how
to succeed. A favorite book to get started on the DevOps journey is The Phoenix Project, which follows the
transformation of a fictional company from NoOps to DevOps. One thing is for certain: DevOps is no longer a "nice
to have" when deploying complex, Cloud Native Applications. It's a requirement and should be planned for and
resourced at the start of any project.
References
Azure DevOps
Azure Resource Manager
Terraform
Azure CLI
PR EVIO U S NE XT
Summary
In summary, here are important conclusions from this guide:

Cloud-native is about designing modern applications that embrace rapid change, large scale, and
resilience, in modern, dynamic environments such as public, private, and hybrid clouds.
The Cloud Native Computing Foundation (CNCF) is an influential open-source consortium of over 300
major corporations. It's responsible for driving the adoption of cloud-native computing across technology
and cloud stacks.
CNCF guidelines recommend that that cloud-native applications embrace six important pillars as shown
in Figure 11-1:
Figure 11-1 . Cloud-native foundational pillars

These cloud-native pillars include:
The cloud and its underlying service model
Modern design principles
Microservices
Containerization and container orchestration
Cloud-based backing services, such as databases and message brokers
Automation, including Infrastructure as Code and code deployment
Kubernetes is the hosting environment of choice for most cloud-native applications. Smaller, simple
services are sometimes hosted in serverless platforms, such as Azure Functions. Among many key
automation features, both environments provide automatic scaling to handle fluctuating workload volumes.
Ser vice communication becomes a significant design decision when constructing a cloud-native
application. Applications typically expose an API gateway to manage front-end client communication. Then
backend microservices strive to communicate with each other implementing asynchronous communication
patterns, when possible.
gRPC is a modern, high-performance framework that evolves the age-old remote procedure call (RPC)
protocol. Cloud-native applications often embrace gRPC to streamline messaging between back-end
services. gRPC uses HTTP/2 for its transport protocol. It can be up to 8x faster than JSON serialization with
message sizes 60-80% smaller. gRPC is open source and managed by the Cloud Native Computing
Foundation (CNCF).
Distributed data is a model often implemented by cloud-native applications. Applications segregate
business functionality into small, independent microservices. Each microservice encapsulates its own
dependencies, data, and state. The classic shared database model evolves into one of many smaller
databases, each aligning with a microservice. When the smoke clears, we emerge with a design that exposes
a database-per-microservice model.
No-SQL databases refer to high-performance, non-relational data stores. They excel in their ease-of-use,
scalability, resilience, and availability characteristics. High volume services that require sub second response
time favor NoSQL datastores. The proliferation of NoSQL technologies for distributed cloud-native systems
can't be overstated.
NewSQL is an emerging database technology that combines the distributed scalability of NoSQL and the
ACID guarantees of a relational database. NewSQL databases target business systems that must process
high-volumes of data, across distributed environments, with full transactional/ACID compliance. The Cloud
Native Computing Foundation (CNCF) features several NewSQL database projects.
Resiliency is the ability of your system to react to failure and still remain functional. Cloud-native systems
embrace distributed architecture where failure is inevitable. Applications must be constructed to respond
elegantly to failure and quickly return to a fully functioning state.
Ser vice meshes are a configurable infrastructure layer with built-in capabilities to handle service
communication and other cross-cutting challenges. They decouple cross-cutting responsibilities from your
business code. These responsibilities move into a service proxy. Referred to as the Sidecar pattern , the
proxy is deployed into a separate process to provide isolation from your business code.
Obser vability is a key design consideration for cloud-native applications. As services are distributed across
a cluster of nodes, centralized logging, monitoring, and alerts, become mandatory. Azure Monitor is a
collection of cloud-based tools designed to provide visibility into the state of your system.
Infrastructure as Code is a widely accepted practice that automates platform provisioning. Your
infrastructure and deployments are automated, consistent, and repeatable. Tools like Azure Resource
Manager, Terraform, and the Azure CLI, enable you to declaratively script the cloud infrastructure you
require.
Code automation is a requirement for cloud-native applications. Modern CI/CD systems help fulfill this
principle. They provide separate build and deployment steps that help ensure consistent and quality code.
The build stage transforms the code into a binary artifact. The release stage picks up the binary artifact,
applies external environment configuration, and deploys it to a specified environment. Azure DevOps and
GitHub are full-featured DevOps environments.
PR EVIO U S

Cloud Native App

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cloud Native App

Uploaded by

Copyright:

Available Formats

Contents

Architecting Cloud-Native .NET Apps for Azure

Rob Vettor , Principal Cloud System Architect/IP Architect - thinkingincloudnative.com, Microsoft

Participants and Reviewers:

Cesar De la Torre , Principal Program Manager, .NET team, Microsoft

Maira Wenzel , Program Manager, .NET team, Microsoft

Who should use this guide

How you can use this guide

Send your feedback

Another day, at the office, working on "the next big thing."

Figure 1-1 . Traditional monolithic design

Figure 1-2 . Cloud-native design

Netflix Has 600+ services in production. Deploys hundred times per

Uber Has 1,000+ services in production. Deploys several thousand

WeChat Has 300+ services in production. Deploys 1,000 times a day.

1 Code Base A single code base for each

2 Dependencies Each microservice isolates and packages

3 Configurations Configuration information is moved out

4 Backing Services Ancillary resources (data stores, caches,

5 Build, Release, Run Each release must enforce a strict

6 Processes Each microservice should execute in its

7 Port Binding Each microservice should be self-

8 Concurrency Services scale out across a large

9 Disposability Service instances should be disposable,

10 Dev/Prod Parity Keep environments across the

11 Logging Treat logs generated by microservices

12 Admin Processes Run administrative/management tasks

13 API First Make everything a service. Assume

14 Telemetry On a workstation, you have deep

15 Authentication/ Authorization Implement identity from the start.

Figure 1-5 . Multiple containers running on a container host

Figure 1-6 . What container orchestrators do

Scheduling Automatically provision container instances.

Affinity/anti-affinity Provision containers nearby or far apart from each other,

Health monitoring Automatically detect and correct failures.

Failover Automatically reprovision failed instance to healthy machines.

Scaling Automatically add or remove container instance to meet

Networking Manage a networking overlay for container communication.

Service Discovery Enable containers to locate each other.

Rolling Upgrades Coordinate incremental upgrades with zero downtime

Figure 1-7 . Common backing services

Figure 1-9 . Azure DevOps offerings

Figure 1-10 . Strategies for migrating legacy workloads

Figure 2-1 . eShopOnContainers Sample App Screenshot.

Features and requirements

Figure 2-2 . eShopOnContainers reference application development architecture.

Overview of the code

Figure 2-5 . The eShopOnContainers Architecture.

Container orchestration and clustering

Azure Kubernetes Service

Deploying to Azure Kubernetes Service using Helm

Figure 2-6 . The eShopOnContainers helm folder.

Azure Dev Spaces

Azure Functions and Logic Apps (Serverless)

Azure App Configuration

Azure Key Vault

Challenges with monolithic deployments

Figure 3-1 . Monolithic architecture.

What are the benefits of containers and orchestrators?

Figure 3-2 . Decomposing a monolithic app to embrace microservices.

What are the scaling benefits?

What scenarios are ideal for containers and orchestrators?

az acr login --name registryname