100% found this document useful (1 vote)

205 views

Beginning Azure Functions

Uploaded by

henchiri.becher

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (1 vote)

205 views

Beginning Azure Functions

Uploaded by

henchiri.becher

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 186

Beginning Azure

Functions
Building Scalable and Serverless Apps
—
Second Edition
—
Rahul Sawhney
Kalyan Chanumolu
Beginning Azure
Functions
Building Scalable
and Serverless Apps
Second Edition

Rahul Sawhney
Kalyan Chanumolu
Beginning Azure Functions: Building Scalable and Serverless Apps
Rahul Sawhney Kalyan Chanumolu
Hyderabad, India Hyderabad, India

ISBN-13 (pbk): 978-1-4842-9202-0 ISBN-13 (electronic): 978-1-4842-9203-7

https://doi.org/10.1007/978-1-4842-9203-7

Copyright © 2023 by Rahul Sawhney and Kalyan Chanumolu

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
Trademarked names, logos, and images may appear in this book. Rather than use a trademark
symbol with every occurrence of a trademarked name, logo, or image we use the names, logos,
and images only in an editorial fashion and to the benefit of the trademark owner, with no
intention of infringement of the trademark.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if
they are not identified as such, is not to be taken as an expression of opinion as to whether or not
they are subject to proprietary rights.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal
responsibility for any errors or omissions that may be made. The publisher makes no warranty,
express or implied, with respect to the material contained herein.
Managing Director, Apress Media LLC: Welmoed Spahr
Acquisitions Editor: Smriti Srivastava
Development Editor: Laura Berendson
Coordinating Editor: Mark Powers
Copyeditor: Kim Burton
Cover designed by eStudioCalamar
Cover image by Hushaan @fromtinyisles on Unsplash (www.unsplash.com)
Distributed to the book trade worldwide by Apress Media, LLC, 1 New York Plaza, New York, NY
10004, U.S.A. Phone 1-800-SPRINGER, fax (201) 348-4505, e-mail [email protected],
or visit www.springeronline.com. Apress Media, LLC is a California LLC and the sole member
(owner) is Springer Science + Business Media Finance Inc (SSBM Finance Inc). SSBM Finance
Inc is a Delaware corporation.
For information on translations, please e-mail [email protected]; for
reprint, paperback, or audio rights, please e-mail [email protected].
Apress titles may be purchased in bulk for academic, corporate, or promotional use. eBook
versions and licenses are also available for most titles. For more information, reference our Print
and eBook Bulk Sales web page at http://www.apress.com/bulk-sales.
Any source code or other supplementary material referenced by the author in this book is
available to readers on GitHub (https://github.com/Apress). For more detailed information,
please visit http://www.apress.com/source-code.
Printed on acid-free paper
Table of Contents
About the Authors��ix

About the Technical Reviewers��xi

Introduction��xiii

Chapter 1: Introduction to Azure Functions��1

Overview of Serverless Computing��2
Overview of Azure Functions��3
Azure Functions Runtime Versions��4
Azure Functions vs. Azure WebJobs��5
Azure Functions vs. Azure Logic Apps��7
Azure Functions Pricing Plans��8
Consumption Plan��9
App Service Plan��9
Premium Plan��10
Summary��12

Chapter 2: Creating Azure Functions��13

Creating an Azure Function Using the Azure Portal��13
Creating an Azure Account��14
Creating Your First Function App Using the Azure Portal��14
Creating Your First Function in the Function App��19
File Hierarchy, Configuration, and Settings in Azure Functions��24
Summary��28

iii
Table of Contents

Chapter 3: Understanding Azure Functions Triggers and Bindings��29

Overview of Triggers and Bindings��29
Register Binding Extensions��32
Binding Expression Patterns��32
Installing Extensions Using Visual Studio Tools��34
Installing Extensions Using the Azure Functions Core Tools��34
Creating an Azure Blob Storage–Triggered Function��35
Creating a Blob-Triggered Function Using C#��35
Blob-Triggered Function Using Node.js��43
Running the Example��49
Summary��49

Chapter 4: Serverless APIs Using Azure Functions��51

Monolithic Architecture vs. Microservice Architecture��52
Converting Monolithic Applications to Highly Scalable APIs Using Azure
Functions��54
Creating an HTTP-Triggered Function with SQL Server Interaction��57
Creating a SQL Server Instance with Sample Data��57
Creating an HTTP-Triggered Function Using C#��61
Creating an HTTP-Triggered OData API for SQL Server Using
Azure Functions��71
Summary��77

Chapter 5: Azure Durable Functions��79

Overview of Durable Functions��79
Types of Functions��80
Application Patterns��81
Function Chaining��81

iv
Table of Contents

Fan-Out/Fan-In��83
Async HTTP APIs��84
Monitoring��86
Human Interaction��87
Bindings for Durable Functions��89
Activity Triggers��89
Orchestration Triggers��90
Orchestration Client��93
Performance and Scaling of Durable Functions��94
Internal Queue Triggers��95
Orchestrator Scale-Out��96
Autoscaling��97
Concurrency Throttling��98
Orchestrator Function Replay��98
Performance Targets��99
Creating Durable Functions Using the Azure Portal��100
Creating a Durable Function��100
Disaster Recovery and Geodistribution of Durable Functions��110
Summary��111

Chapter 6: Deploying Functions to Azure��113

Deploying Functions Using Continuous Deployment��113
Setting Up a Code Repository for Continuous Deployment��114
Setting Up an Azure DevOps Account��115
Enabling Deployment Slots��120
Setting up Continuous Deployment for Azure Functions��122
Deploying Azure Functions Using ARM Templates��127

v
Table of Contents

Deploying a Function App on the Consumption Plan��128

Deploying a Function App on the App Service Plan��135
Summary��144

Chapter 7: Getting Functions Production-Ready��145

Using Built-in Logging��145
Using Application Insights to Monitor Azure Functions��147
Application Insights Settings for Azure Functions��147
Integrate Application Insights During New Azure Function Creation��147
Manually Connecting Application Insights to Azure Functions��150
Disabling Built-in Logging��154
Configuring Categories and Log Levels��154
Securing Azure Functions��156
Configuring CORS on Azure Functions��161
Summary��163

Chapter 8: Best Practices for Working with Azure Functions��165

Organize Your Functions��165
Based on Business Function��165
For Performance and Scaling��166
By Configuration and Deployment��166
By Privilege��167
By App Visibility��167
Separate Function Apps for Each Function��167
Develop Robust Functions��168
Design for Statelessness��168
Use Messaging Services��168
Integrate with Monitoring Tools��169

vi
Table of Contents

Use Separate Storage Accounts��169

Account for Delays with Function Cold Starts��169
Use Asynchronous Patterns��170
Manage Costs��171
Summary��171

Index��173

vii
About the Authors
Rahul Sawhney works as a software developer
with Microsoft, India. He has more than five
years of experience delivering cloud solutions
using technologies such as .NET Core, Azure
Functions, microservices, AngularJS, Web
API, Azure AD, Azure Storage, ARM templates,
Azure App Service, Azure Traffic Manager,
and more. He is a Microsoft Certified Azure
Developer and Architect. He loves learning
new technologies and is passionate about Microsoft technologies. In his
free time, he loves playing table tennis, watching movies, and reading
books.

Kalyan Chanumolu is a Senior Technical

Program Manager at Microsoft. He works on
building the engineering systems that power
Azure—the world’s computer. He is extremely
passionate about distributed systems and
cloud computing and has served as a technical
reviewer for books on microservices, Azure,
ASP.NET, and more. He has vast experience
in software development, consulting, and
migrating enterprise workloads to the cloud. He loves cycling, swimming,
and reading books.

ix
About the Technical
Reviewers
Carsten Thomsen is a back-end developer primarily but works with
smaller front-end bits as well. He has authored and reviewed several books
and created numerous Microsoft Learning courses, all related to software
development. He works as a freelancer/contractor in various European
countries; Azure, Visual Studio, Azure DevOps, and GitHub are some of
the tools he regularly uses. He is an exceptional troubleshooter and enjoys
working with architecture, research, analysis, development, testing, and
bug fixing. Carsten is a great communicator with mentoring and team-lead
skills and enjoys researching and presenting new material.

Kapil Bansal is a lead DevOps engineer at S&P Global Market Intelligence,

India. He has more than 14 years of experience in the IT industry, having
worked on Azure cloud computing (PaaS, IaaS, and SaaS), Azure Stack,
DevSecOps, Kubernetes, Terraform, Office 365, SharePoint, release
management, application lifecycle management (ALM), Information
Technology Infrastructure Library (ITIL), and Six Sigma. He has worked
with companies such as IBM India Pvt Ltd, HCL Technologies, NIIT
Technologies, Encore Capital Group, and Xavient Software Solutions,
Noida. He has served multiple clients based in the United States, the
United Kingdom, and Africa, such as T-Mobile, World Bank Group, H&M,
WBMI, Encore Capital, and Bharti Airtel (India and Africa).

xi
Introduction
Get ready to create highly scalable apps and monitor functions in
production using Azure Functions 2.0!
We’ll start by taking you through the basics of serverless technology
and Azure Functions and then cover the different pricing plans of Azure
Functions. After that, we’ll dive into how to use Azure Functions as a
serverless API. We will then walk you through the Durable Functions
model, disaster recovery, and georeplication.
Moving on, we provide practical recipes for creating different functions
in Azure Functions using the Azure portal and Visual Studio Code. Finally,
we discuss DevOps strategy and how to deploy Azure Functions and get
Azure Functions production-ready.
By the end of this book, you will have all the skills needed to work with
Azure Functions, including working with Durable Functions and deploying
functions and making them production-ready by using telemetry and
authentication/authorization.

What This Book Covers

Chapter 1 goes through the basics of serverless computing and talks
about Azure Functions. It compares Azure Functions to WebJobs, so you
understand their differences. We also talk about the different pricing plans
of Azure Functions.
In Chapter 2, you create the first function using the Azure portal
and then Visual Studio Code. We also discuss the Azure Functions file
hierarchy, configuration, and settings.

xiii
Introduction

In Chapter 3, you learn about triggers and bindings. We also discuss

changes to Azure Functions 2.0 bindings, and you create Azure Blob
Storage–triggered functions.
Chapter 4 goes through the differences between monolithic
applications vs. microservices. Then, we talk about converting a
monolithic application to microservices using Azure Functions. You create
some functions and then learn about proxies.
Chapter 5 starts with an overview of the Durable Functions pattern
and bindings. You also learn about performance and scaling in a durable
function. You create your first durable function and learn about disaster
recovery and georeplication.
Chapter 6 looks at deploying functions to Azure, using a CI/CD
pipeline and ARM templates.
Chapter 7 covers the built-in logging capabilities of Azure Functions
and looks at Application Insights and how it is used to monitor Azure
Functions. We also talk about securing Azure Functions using Azure Active
Directory and how to configure cross-origin site scripting (CORS) in Azure
Functions.

Source Code
The source code for this book can be downloaded from github.com/
apress/beginning-azure-functions-2e.
Let’s get started!

xiv
CHAPTER 1

Introduction to Azure
Functions
The software industry is now in an era where apps and services
increasingly rely on the cloud. Applications are globally available and need
to scale at a moment’s notice. Deployments need to be seamless without
any perceivable downtime to the end customer. Periodic maintenance
windows for making infrastructure upgrades are a thing of the past.
To help developers be more productive and focus on developing
software rather than worrying about provisioning and maintaining servers,
and orchestrating deployments, cloud platforms such as Microsoft Azure,
Amazon Web Services, and Google Cloud Platform, provide serverless
compute offerings.
Azure Functions is one such product for serverless computing.
With Azure Functions, developers can write less code and manage less
infrastructure, and organizations can bring applications to market quickly
with fewer upfront costs. Before diving into Azure Functions, let’s discuss
serverless computing and what it means.
This chapter covers the following topics.

• Overview of serverless computing

• Overview of Azure Functions and runtime versions

• Azure Functions vs. Azure WebJobs

• Azure Functions pricing options

© Rahul Sawhney and Kalyan Chanumolu 2023 1
R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_1
Chapter 1 Introduction to Azure Functions

Overview of Serverless Computing

Serverless computing, also known as function as a service (FaaS), is an
event-driven application design and deployment model that allows
developers to build and run applications without managing servers.
Serverless computing does not mean your code runs without a server; it
means you don’t have to take care of the server maintenance, including
patching, upgrading, and so on. The servers are managed by a cloud
service provider such as Amazon, Microsoft, or Google, and you have to
manage your code/application. With serverless computing, you pay only
for the time your code runs or executes. Also, the cloud service provider
takes care of scaling and load balancing, which is a win-win situation
for both the cloud service provider and you because you can dedicate
much of your time to doing what’s most important: developing the code/
application. The cloud service provider maintains and owns the server and
charges you for the usage.
Serverless computing is a paradigm shift in computing. Deploying
applications on physical machines used to take months; with serverless
computing, deploying takes just a few minutes or less.
Figure 1-1 shows how the underlying infrastructure is abstracted from
the developer.

2
Chapter 1 Introduction to Azure Functions

ServerIess
Unit of ScaIe 002D Function
Infrastructure Abstraction

Containers
Unit of ScaIe 002D AppIication

VirtuaI Machines
Unit of ScaIe 002D Machine

PhysicaI Servers
Unit of ScaIe 002D Server

Focus on Business Logic

Figure 1-1. Infrastructure abstraction in cloud computing

Overview of Azure Functions

With Azure Functions, you can start writing code for your application
without worrying about the infrastructure required to run and scale
the application. Azure Functions also provides the capability to scale as
needed. So, if the load is high, you can expect Azure Functions to scale and
cater to the high load. Also, with Azure Functions, you pay only for the time
your code runs, so if the load on the application is low, you pay less.
The following are some important Azure Functions features.

• Browser-based interface: You can write and test your

code directly in the Azure portal without using any
integrated development environment (IDE).

3
Chapter 1 Introduction to Azure Functions

• Programming languages: Azure Functions supports

many languages, such as C#, JavaScript, F#, Java,
Python, TypeScript, PHP, Batch, Bash, PowerShell, and
a few other experimental languages.

• Seamless integration with third-party apps: Azure

Functions integrates seamlessly with third-party apps
such as Facebook, Google, Twitter, and Twilio, and
other Azure services, like Cosmos DB, Azure Storage,
Azure Service Bus, and more. You can also integrate
existing apps using triggers and events.

• Continuous deployment: Azure Functions supports

continuous deployment through Azure DevOps (VSTS),
GitHub, Xcode, Eclipse, and IntelliJ IDEA.

As you can see, Azure Functions possesses some unique capabilities

that enhance your productivity and provide lots of different options for
developers to choose from. Still, we have heard developers becoming
confused about when to use Azure Functions and when to use Azure
WebJobs. The primary reason for this confusion is that developers have
traditionally reduced the load on the application by doing extensive and
time-consuming computations in Azure WebJobs.
The next section discusses the differences between Azure Functions
and Azure WebJobs and in which scenario you should use each.

Azure Functions Runtime Versions

The Azure Functions runtime provides a hosting platform for functions
written in many different languages and supports a wide variety of triggers
and bindings. Support for new languages and integration with other cloud
services are distributed as updates to the runtime.

4
Chapter 1 Introduction to Azure Functions

Azure Functions currently has four versions of the runtime host.

Table 1-1 describes when these versions should be used.

Table 1-1. Azure Functions Runtime Host versions

Version Description

4.x The latest and recommended runtime version for functions in all
languages. This version supports running C# functions on .NET 6.0 and
above.
3.x This runtime version supports all languages. This version supports
running C# functions on .NET Core 3.1 and .NET 5.0.
2.x Currently, this version is in maintenance mode. All the enhancements
are provided in newer versions only. This version supports running C#
functions on .NET Core 2.1
1.x This version is in maintenance mode and should only be used for C#
apps that must use .NET Framework. It only supports development
locally on Windows, the Azure Stack Hub portal, and the Azure portal.

Going forward in this book, functions are created using version 4.x.

Azure Functions vs. Azure WebJobs

Azure Functions and Azure WebJobs are code-first integration services
designed for developers. Both support features such as authentication,
Application Insights, and source control integration. Table 1-2 notes some
differences between Azure Functions and WebJobs.

5
Chapter 1 Introduction to Azure Functions

Table 1-2. Azure Functions and WebJobs Differences

Azure Functions Azure WebJobs

Trigger Azure Functions can be triggered with There are two types of
many triggers, as discussed in the WebJobs: triggered and
upcoming chapters. Azure Functions continuous.
don’t run continuously.
Supported Azure Functions support the following Azure Web Jobs also
Languages languages: C#, Java, JavaScript, supports languages like
PowerShell, Python, and TypeScript. JavaScript, C#, and F#.
Deployment Azure Functions run using a classic\ It runs as a background
dynamic App Service plan. process in app services,
such as API apps, mobile
apps, and web apps.

Azure Functions is a great choice in most scenarios because it offers

many programming languages, flexible pricing options, and increased
developer productivity. However, the following are two scenarios where
you should use WebJobs instead.

• You have an Azure App Service environment where you

want to run some code snippets and maintain the same
DevOps pipeline and environment.

• You need to customize the behavior of the host that

listens for triggers and calls functions, such as having
custom retry policies or error handling.

Azure WebJobs runs under the Azure App Service model, whereas
Azure Functions has different pricing models that give you more control
over pricing. You’ll learn about pricing next.

6
Chapter 1 Introduction to Azure Functions

Azure Functions vs. Azure Logic Apps

Azure Logic Apps is a designer-first low-code/no-code cloud platform
where you can create and run automated workflows using a visual designer
and selecting operations from prebuilt connectors. You can quickly build
workflows by dragging and dropping API connectors onto the Logic Apps
designer, specify triggers, configure inputs/outputs and connect one action
to another. All from a user interface without having to write complex code
to integrate with other Azure services. Table 1-3 notes some differences
between Azure Functions and Logic Apps.

Table 1-3. Azure Functions Logic Apps Differences

Azure Functions Logic Apps

Trigger An Azure Function can have only Azure Logic Apps can have up
one trigger. to ten triggers.
Supported Azure Functions supports the The source code generated by
Languages following languages: C#, Java, the graphic designer can be
JavaScript, PowerShell, Python, and accessed as a JSON template.
TypeScript.
Development Azure Functions can be developed Logic Apps are authored using
on the Azure portal using Visual the Azure portal or Visual Studio
Studio or any supported IDE and using an extension and produce
have support for debugging. They a JSON template that can be
can be deployed using Visual checked into source control and
Studio, Azure Pipelines, or GitHub deployed using Azure DevOps,
Actions. Visual Studio, or FTP.
(continued)

7
Chapter 1 Introduction to Azure Functions

Table 1-3. (continued)

Azure Functions Logic Apps

Hosting Azure Functions runs using a Logic Apps run in a Standard

Plans classic\dynamic App Service plan. plan (Single tenant)/
Consumption plan (Multitenant).
Billing Billing is based on resource Billing is based on the number
consumption and executions per of times triggers, actions, or
second. connectors are executed.

Logic Apps are especially useful when you coordinate simple actions
across multiple systems and services to automate business processes. They
provide rich capabilities for control flow operations such as conditions,
switches, loops, and scopes. You can also implement error and exception
handling in your workflows. Azure Functions are better suited for
workflows with complex business logic and data transformations, reuse of
components and libraries across other projects and scenarios that require
more control on security, reliability, and availability. Azure Functions and
Logic Apps can call each other, allowing you to design a solution that can
take advantage of both technologies as appropriate for your scenario.

Azure Functions Pricing Plans

Azure Functions supports three pricing plans.

• Consumption plan

• App Service plan

• Premium plan
Let’s look at each plan in detail.

8
Chapter 1 Introduction to Azure Functions

Consumption Plan
Azure Functions’ Consumption plan allows you to pay only when your
code is executing. This helps you save significantly over the App Service
plan or when using a virtual machine. For example, if you have a weekly
newsletter for your website, instead of using WebJobs, you can use Azure
Functions and save a significant amount of money.
The metric for calculating price in Azure Functions is gigabytes/
seconds (GB/s). This metric calculates the memory usage and total
execution time for billing. It is billed based on per-second resource
executions and consumptions.
The Consumption plan grants one million requests and 400,000 GB/s
of resource consumption for free per month per subscription across all
Azure Functions apps in that subscription.

App Service Plan

Azure Functions’ App Service plan is the same plan used for hosting a
website, the Web API, and so on. With the App Service plan for Azure
Functions, instead of paying for the duration when a function is executing,
you pay for the reserved resources of the underlying virtual machine (VM).
This makes the App Service plan costlier than the Consumption plan.
Why do some companies use the App Service plan? The reason is that
in the Consumption plan, functions have a time limit of five minutes, so
if your Azure Functions code runs for more than five minutes in a single
execution, it times out. In contrast, there is no time limitation for Azure
Functions in the App Service plan. So, in the App Service plan, Azure
Functions is as good as WebJobs.
Also, when you are billed on the App Service plan, it is easier to
maintain the monthly quotas of your company because all the resources
are under the same App Service plan.

9
Chapter 1 Introduction to Azure Functions

However, if you have a piece of code that is resource hungry, having

it on the same App Service plan as the rest of your company would make
your other applications vulnerable because Azure Functions would
be using the same shared resources and thus would make the other
applications run slower.

Premium Plan
Azure Functions’ Premium plan is a dynamic scale hosting plan for
function apps. Azure Functions Premium plan offers the following
benefits.

• More predictable pricing compared to the

Consumption plan

• Premium instance sizes, such as one, two, and four core

instances

• Unlimited execution duration, with 60 minutes

guaranteed

• Virtual network connectivity

• Perpetually warm instances, eliminating one of the

biggest serverless problems of cold start

With the Premium plan, there is no execution charge. The billing is

based on the number of cores and memory allocated across instances,
whereas in the Consumption plan, you are billed per execution and
memory used.
In the Premium plan, at least one instance must be allocated at all
times. Also, all function apps in a Premium plan share allocated instances.
This results in a minimum monthly cost per active plan regardless of
whether the function is alive or idle.

10
Chapter 1 Introduction to Azure Functions

How the Premium Plan Solves a Cold Start Problem

A cold start refers to a phenomenon where an application that isn’t used
for a while takes longer to start up. In Azure Functions, latency is the time a
user must wait for their function to execute.
In Azure Functions, a cold start results in higher latency for a function
that hasn’t been called recently. Let’s now look at what happens under the
hood when you write a function.

• Azure allocates a server with function and a server with

capacity.

• The function runtime then starts up on that host.

• Your code then executes.

In the Consumption plan, if an event or execution doesn’t occur,

your app may scale down to zero instances. So, after that, the first call
the function receives goes through the three steps and results in higher
latency for the first few calls. Figure 1-2 shows how a cold start differs from
a warm start.

Figure 1-2. Cold start vs. warm start

11
Chapter 1 Introduction to Azure Functions

On the other hand, in the Premium plan, there are two ways to solve
a cold start problem. First, always ready instances and pre-warmed
instances. The Premium plan lets you configure up to 20 instances as
always ready instances. When an event triggers a function app, the
execution is routed to an always ready instance. Additional instances are
warmed as a buffer as the function gets more traffic. These buffer instances
are then used during scale to prevent a cold start. These buffer instances
are called pre-warmed instances.

Summary
You should now have a basic understanding of serverless computing,
Azure Functions, and app pricing. Chapter 2 builds upon this chapter and
gets started with creating Azure Functions.

12
CHAPTER 2

Creating Azure
Functions
In this chapter, you will learn how to create an Azure function using the
Azure portal and Visual Studio Code. You will also gain insight into a
function’s file hierarchy, configuration, and settings.
In this chapter, you start using Azure Functions. The latest runtime
version is 4.0, which is used throughout the book’s examples. Over the
course of this chapter, we will cover the following topics:

• Creating an Azure portal account

• Creating functions using the Azure portal

• Azure Functions file hierarchy, configuration, and

settings

Let’s examine the two ways you can create functions with Azure
Functions.

reating an Azure Function Using

C
the Azure Portal
In this section, you will create your first function using the Azure portal: a
boilerplate application.

© Rahul Sawhney and Kalyan Chanumolu 2023 13

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_2
Chapter 2 Creating Azure Functions

Creating an Azure Account

First, you need to log in to the Azure portal. You can go to http://portal.
azure.com to sign in. If you don’t have an Azure subscription or are a first-
time Azure user, you can get an Azure account for free for 12 months. Visit
https://azure.microsoft.com/en-us/free/ to get started.
To get started with Azure Functions, visit https://functions.azure.
com/ and click the Free Account button.

reating Your First Function App Using

C
the Azure Portal
Once you have completed the sign-up/sign-in process, you are taken to the
Azure portal dashboard. Now you can create a function app.
On the dashboard, click “Create a resource” in the left panel. In the
menu, click Compute, search for “Function”, and select Function App, as
shown in Figure 2-1.

Figure 2-1. Creating a function app resource

14
Chapter 2 Creating Azure Functions

You are asked to provide details such as the app name, resource group,
OS, hosting plan, and so on. Refer to Figure 2-2 to enter the settings.

• Function App Name: This is the name of the

function app. The app name in this example is
BuildingAzureFunction, so the URI to access
the function app is BuildingAzureFunction.
azurewebsites.net.

• Subscription: This is the subscription under which

the function app is created. An Azure account can
have multiple subscriptions used for maintenance and
billing purposes.

• Publish: There are two options to publish this function:

Code and Docker Container. In Code Publish, you
can choose among the Consumption, App Service,
or Premium plans; but the Consumption plan is not
available in Docker Container.

• Plan Type: By default, the Consumption plan is

selected. This plan is ideal for most scenarios, and you
are billed only for the duration your function executes.
But you can choose the App Service plan or Premium
as well.

• Region: Always choose the location nearest to where

you expect the majority of the traffic to come for your
function app.

• Storage: Every function app requires storage for

logging function executions and managing schedules
for function triggers. To use an existing storage account,
go to Hosting and select Azure Storage Account. By
default, it creates a new storage account.

15
Chapter 2 Creating Azure Functions

• Operating System: Azure Functions supports both

Linux and Windows. You can select either of them
based on your preferred runtime stack.

Figure 2-2. Create function app

16
Chapter 2 Creating Azure Functions

Click the Review + Create button to get the summary of the function,
and then click the Create button, as shown in Figure 2-3.

Figure 2-3. Summary view of the function app

17
Chapter 2 Creating Azure Functions

You can check the status of your function by clicking the bell icon at
the top, as shown in Figure 2-4.

Figure 2-4. Checking the status

Once the deployment is completed, your first Azure Functions app,

named BuildingAzureFunction, is ready, and you can now start coding for
it. To go to the function app you created in the previous steps, navigate to
the resource group. Please refer to Figure 2-5.

Figure 2-5. Select the resource group you created

18
Chapter 2 Creating Azure Functions

reating Your First Function

C
in the Function App
Your function app is ready and deployed, but it is not usable because you
haven’t created any functions yet. A function app is like a container that
can hold multiple functions. A function is the component of the function
app where you code your logic.
To create your first function, click BuildingAzureFunction, as
shown in Figure 2-5. Then click the Function menu inside the Functions
subdirectory, and click Create, as shown in Figure 2-6.

Figure 2-6. Create function app

For this example, choose “Develop in portal”. However, you can use
VS Code or any other editor that you are comfortable with. Several trigger
templates can trigger your Azure Function. Select the HTTP trigger
template for this example and provide your function app’s details and the
authorization level, as shown in Figure 2-7.

19
Chapter 2 Creating Azure Functions

Figure 2-7. Create HTTP triggered function

The following describes the three authorization levels.

• Function: The authorization level function uses the

function key. The function key can be found on the
Keys Management panel on the portal.

20
Chapter 2 Creating Azure Functions

• Admin: Admin authorization level uses the master

key. Again the master key can be found in the Keys
management panel on the portal.

• Anonymous: No API key is needed to access the

function.

Click the Create button, and your first HTTP trigger function is created.
It comes with basic boilerplate code. This code should be good enough
for you to test your function. Click the Test/Run button, as shown in
Figure 2-8.

Figure 2-8. HTTP Function boilerplate code

21
Chapter 2 Creating Azure Functions

In the Test panel, select Key as the default (Function key), update the
body parameter as needed, and click Run, as shown in Figure 2-9.

Figure 2-9. Running your Azure Function

22
Chapter 2 Creating Azure Functions

The output is shown in the Output tab, as seen in Figure 2-10.

Figure 2-10. Azure Function output

After validating that the function is running as expected, copy the

function URL, as shown in Figure 2-11.

Figure 2-11. Copy function URL

Create another HTTP trigger function, HttpTrigger2, within the same

function app.
You have created your first running Azure Functions app.
Next, let’s look at file hierarchy, configuration, and settings in Azure
Function.

23
Chapter 2 Creating Azure Functions

F ile Hierarchy, Configuration, and Settings

in Azure Functions
The code for all the functions in a specific function app is located in
the /wwwroot/ folder. The hierarchy of the function created is shown in
Figure 2-12.

Figure 2-12. File hierarchy

Note All functions in a function app must share the same language

stack in version 2.x and higher of the Azure Functions runtime.

You can also view the hierarchy of your function. Go to the

Development Tools menu, select Advanced Tools, and click the Go
button, as shown in Figure 2-13. This opens Kudu, an engine behind git
deployments in Azure. You can read more about it at https://github.
com/projectkudu/kudu/wiki.

24
Chapter 2 Creating Azure Functions

Figure 2-13. Advanced Tools in Functions

In the Kudu back end, you can now navigate to Tools ➤ Zip Push
Deploy to see the file hierarchy, as shown in Figure 2-14.

Figure 2-14. File hierarchy in Kudu

The host.json file is at the root of the function app folder and contains
configuration options that affect all functions in a function app instance.
In this file, you can specify the extension bundle version and the default
FunctionTimeout.
The function.json file is inside each function folder and defines the
function’s trigger, bindings, and other configuration settings. The runtime
uses this file to determine the events to monitor and how to pass data into
and return data from a function execution. For compiled languages like C#,
this file is generated automatically. However, for scripting languages like
JavaScript and PowerShell, you must author this file yourself.

25
Chapter 2 Creating Azure Functions

All the application-level extensions, such as Cosmos DB, HTTP

triggers, queues, and so on, reside under the extensions object and not in
the root of the function.json object, as shown in Figure 2-15.

Figure 2-15. Application-level extensions

All the logging-level settings for Azure Functions reside under the
logging object, as shown in Figure 2-16.

26
Chapter 2 Creating Azure Functions

Figure 2-16. Logging-level extensions

27
Chapter 2 Creating Azure Functions

Summary
In this chapter, you created your first function app and function. Also,
you learned about the Azure Functions file hierarchy, configuration,
and settings. In the next chapter, you’ll learn how functions are triggered
and how other Azure services can be declaratively connected to Azure
Functions.

28
CHAPTER 3

Understanding Azure
Functions Triggers
and Bindings
This chapter teaches you two main components of Azure Functions:
triggers and bindings.
This chapter covers the following topics.

• Overview of triggers and bindings

• Register binding extensions

• Binding expression patterns

• Creating an Azure Blob Storage–triggered function

Overview of Triggers and Bindings

A trigger defines the external event that can invoke the Azure function.
Each function can have only one trigger. Triggers usually carry data
associated with the event, which is the payload that triggers the function.

© Rahul Sawhney and Kalyan Chanumolu 2023 29

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_3
Chapter 3 Understanding Azure Functions Triggers and Bindings

Azure Functions support several triggers, and Microsoft constantly

adds support for more services. You can write a custom trigger of your own
if you wish to. The following describes some common triggers.

• HTTPTrigger: This trigger is fired when the Azure

function detects an incoming HTTP request. The request
body and query parameters are provided as input
parameters to the function. In Chapter 2, you created an
HTTP-triggered function using Visual Studio Code.

• BlobTrigger: This trigger gets fired when a blob is

created or updated on an Azure storage account or a
blob. The contents of the blob are provided as an input
parameter to the function.

• QueueTrigger: This trigger gets fired when a new

message arrives in Azure Queue Storage.

• EventHubTrigger: This trigger gets fired when an event

is delivered to the Azure Event Hubs event stream.

• TimerTrigger: This trigger is called on a scheduled

basis. You can set the time to execute the function using
this trigger.
• Service Bus Trigger: This trigger gets fired when a new
message enters an Azure Service Bus topic or queue.

• GitHub Webhook: This trigger gets fired when any

event, such as Create Branch, Delete Branch, Issue
Comment, or Commit Comment, occurs in your
GitHub repository.

Azure Functions bindings are a declarative way of connecting another

resource to a function. Bindings can be connected as input bindings,
output bindings, or both. Data from these bindings is provided to the
function as parameters.

30
Chapter 3 Understanding Azure Functions Triggers and Bindings

Azure Functions has the following supported bindings, and the list is
ever-growing.

• Blob Storage

• Cosmos DB

• Event Grid

• Event Hubs

• HTTP and webhooks

• Queue Storage

• Table Storage

• Service Bus

• SendGrid

• SignalR

• Twilio

Bindings are optional in Azure Functions, and you can have multiple
input and output bindings. Azure Functions triggers and bindings are
configured in the functions.json file, and they help you avoid putting
hard-coded values in the code.

Note You can find all the supported bindings in Azure Functions by
visiting https://docs.microsoft.com/en-us/azure/azure-
functions/functions-triggers-bindings#supported-
bindings.

31
Chapter 3 Understanding Azure Functions Triggers and Bindings

Register Binding Extensions

Extensions are a way of creating and distributing triggers and bindings for
Azure Functions. In the Azure Functions 2.0 runtime, a new binding model
was introduced where the runtime is decoupled from the extensions
except for a few core bindings, like HTTP and timer. It provides additional
flexibility and reduces the load time by loading only the extensions
referenced in the function app. It also means that the runtime has no
knowledge of the extensions, so they must be registered before use.
Extensions are distributed as NuGet packages, and these extensions
are registered by installing the required NuGet package for the extension.

Binding Expression Patterns

Binding Expressions is a very useful capability of triggers and bindings. In
your function code, function parameters, and function.json file, you can
use expressions that resolve to values from various sources during runtime
instead of specifying values during compilation or deployment time.
For example, you have a function with blob output binding, and you
need to create a file with the current date and time as the file name. You
could use the {DateTime} binding expression in the function.json file, as
follows.

{
  "type": "blob",
  "name": "blobOutput",
  "direction": "out",
  "path": "my-output-container/{DateTime}.txt"
}

32
Chapter 3 Understanding Azure Functions Triggers and Bindings

Let’s look at another example of using function app settings in binding

expressions. Assume that you have a timer-triggered function, and you do
not want to hardcode the timer schedule in your code. You could create a
new app setting called ScheduleConfiguration to specify the schedule and
reference the app settings, as follows. App setting binding expressions are
wrapped in percent signs rather than curly braces.

[Function("RecurringOrdersProcessor")]
public async Task Run([TimerTrigger("%ScheduleConfiguration%",
RunOnStartup = false, UseMonitor = false)] MyInfo myTimer)
{

// Function App code

The following are types of binding expressions.

• App settings

• Trigger file name

• Trigger metadata

• JSON payloads
• New GUID

• Current date and time

Note You can find all the supported binding expressions by visiting
https://learn.microsoft.com/en-us/azure/azure-
functions/functions-bindings-expressions-patterns.

33
Chapter 3 Understanding Azure Functions Triggers and Bindings

Installing Extensions Using Visual

Studio Tools
With Visual Studio Code or Visual Studio, you are referencing the
extensions package directly from the project. The tool handles the
installation and registration of the extensions.
Let’s look at an example of installing an extension.
In the Visual Studio Code terminal window, run the following
command to install and register the Azure Blob Storage bindings and
trigger to your function.

Dotnet add package Microsoft.Azure.Functions.Worker.Extensions.

Storage.Blobs --version 5.0.0

Installing Extensions Using the Azure

Functions Core Tools
Azure Functions Core Tools lets you develop and test your functions on
your local computer from the command prompt or terminal. It is available
as a download for Windows, macOS, and Linux. You can download and
install it on your developer machine and use the func extension install
command for managing extensions. This command lets you install an
extension and register it to the function.
The following is an example of installing an extension.

func extensions install –package Microsoft.Azure.Functions.

Worker.Extensions.Storage.Blobs –version 5.0.0

This command installs the blob extension to Azure Functions, allowing

you to configure your function for a blob trigger.

34
Chapter 3 Understanding Azure Functions Triggers and Bindings

Note With any of the installation steps mentioned, if you install and
register the extension, a metadata file named extensions.json is
generated in the bin folder inside the function’s app root folder. The
runtime uses only the extensions registered in this file.

reating an Azure Blob Storage–

C
Triggered Function
This section teaches you how to create an Azure Blob Storage–triggered
function using C# and Node.js. We will walk you through each process
step-by-step. Visual Studio Code is used to create the function. To set up
the system to create a function, review the “Creating an Azure Function
Using Visual Studio Code” section in Chapter 2.
This function resizes the image once uploaded to the blob by using a
blob trigger in Azure Functions.
Let’s start by first creating a function using C#.

reating a Blob-Triggered Function

C
Using C#
Set up your machine as covered in Chapter 2. Once the machine is set up,
open Visual Studio Code, go to the Extensions section, and install the latest
versions of Azure Functions and C# extensions. Once this is done, go to the
Azure Functions menu and add a new function with the following steps.
Open Visual Studio Code and click the Azure logo in the left menu, as
shown in Figure 3-1.
Click the plus icon and create a new function, as shown in Figure 3-2.

35
Chapter 3 Understanding Azure Functions Triggers and Bindings

Figure 3-1. Clicking the Azure logo

Figure 3-2. Creating a new function app

Create a new project on the next dialog and select a folder to save the
project files.
Select the language in which you want to code your function. In this
example, C# was selected, as shown in Figure 3-3.

Figure 3-3. Selecting the language

36
Chapter 3 Understanding Azure Functions Triggers and Bindings

Select the .NET 6.0 Isolated LTS runtime version, as shown in

Figure 3-4. (This programming model was discussed in Chapter 2.)

Figure 3-4. Selecting the .NET runtime

Select “Azure Blob Storage trigger” as the project template, as shown in

Figure 3-5.

Figure 3-5. Selecting the template for the function trigger

Provide a name for your function app, as shown in Figure 3-6.

Figure 3-6. Naming the function

Provide the namespace. By default, it is Company.Function. For this

demo, set it as AzureFunctionV2Book.Function, as shown in Figure 3-7.
37
Chapter 3 Understanding Azure Functions Triggers and Bindings

Figure 3-7. Providing the namespace

Click “Create new local app setting,” as shown in Figure 3-8.

Figure 3-8. Creating a new local app setting

The next screen asks you to sign in to Azure. Select “Sign in to Azure”
if you are not signed in. All the subscriptions associated with your
Azure account load. Select the subscription, as shown in Figure 3-9. The
subscription name in this example is VSEng.

Figure 3-9. Selecting the subscription

You can select the Azure storage account that already exists or create a
new one. In this case, you select the existing one, as shown in Figure 3-10.

Figure 3-10. Selecting an existing storage

38
Chapter 3 Understanding Azure Functions Triggers and Bindings

Once the storage account is set up, provide a name for the blob trigger.
This is the path that the trigger monitors. By default, it shows as samples-
workitems. For this function, you are setting it to function-v4-book, as
shown in Figure 3-11.

Figure 3-11. Setting it to function-v4-book

If you get a prompt, as shown in Figure 3-12, select “Use local

emulator” and continue.

Figure 3-12. Selecting storage account for Azure Functions runtime

Add your project to a workspace in Visual Studio Code for easy access,
as shown in Figure 3-13.

Figure 3-13. Adding the function to a workspace

Your function is set up, as shown in Figure 3-14.

39
Chapter 3 Understanding Azure Functions Triggers and Bindings

Figure 3-14. Completed function

Now you need to add a package named SixLabors.ImageSharp to the

project. To do that, type dotnet add package SixLabors.ImageSharp in
the terminal. Visual Studio Code installs the package.
Once the package is installed, paste the following code into the .cs file.

using Microsoft.Azure.Functions.Worker;
using Microsoft.Extensions.Logging;
using SixLabors.ImageSharp;
using SixLabors.ImageSharp.Formats;
using SixLabors.ImageSharp.Processing;

namespace AzureFunctionV4Book.Function
{
    public class MyFirstBlobTriggerFuntionApp
    {
        private readonly ILogger _logger;

40
Chapter 3 Understanding Azure Functions Triggers and Bindings

        public MyFirstBlobTriggerFuntionApp(ILoggerFactory
loggerFactory)
        {
            _logger = loggerFactory.CreateLogger<MyFirstBlobTri
ggerFuntionApp>();
        }

        [Function("MyFirstBlobTriggerFuntionApp")]
        [BlobOutput("output-blob/{name}_resized", Connection
= "functionv4book_STORAGE")] // Output Container and
connection string of the storage account
        public byte[] Run([BlobTrigger("image-blob/{name}",
Connection = "functionv4book_STORAGE")] byte[]
inputImage, string name)
        {
            _logger.LogInformation($"C# Blob trigger function
processed blob\n Name:{name} \n Size: {inputImage.
Length} Bytes");

            IImageFormat format;
            var width = 100;
            var height = 200;

using (Image input = Image.Load(inputImage, out

format))
{
var output = new MemoryStream();

input.Mutate(x => x.Resize(width, height));

                input.Save(output, format);
                return output.ToArray();
            }

41
Chapter 3 Understanding Azure Functions Triggers and Bindings

}

}
}

In the top menu, click Debug ➤ Start Without Debugging to get

the function up and running. Now, go to Azure Storage and create two
containers named image-blob and output-blob.
If you have never created a container, go to https://docs.microsoft.
com/en-us/azure/storage/blobs/storage-quickstart-blobs-portal to
create containers.
Once the container is created, upload the blob as shown in the
previous link in the image-blob container. You see the function being
triggered. Once the function runs, the resized image appears in output-
blob, as shown in Figure 3-15a and Figure 3-15b.

Figure 3-15a. File size before resizing the image

42
Chapter 3 Understanding Azure Functions Triggers and Bindings

Figure 3-15b. File sizes after resizing the image

You have created a blob-triggered function using C#.

Next, let’s look at creating a blob-triggered function using Node.js.

Blob-Triggered Function Using Node.js

Let’s implement the same functionality of image resizing using Node.js.
Set up the machine for Node.js by installing the latest version of Node.
js. Once that is done, restart Visual Studio Code and go to the function. The
first two steps are the same as what you did while creating a blob-triggered
function using C#.
Select JavaScript as the language, as shown in Figure 3-16.

Figure 3-16. Selecting the language

43
Chapter 3 Understanding Azure Functions Triggers and Bindings

Select the “Azure Blob Storage trigger” template, as shown in

Figure 3-17.

Figure 3-17. Selecting the template

Provide the function name. By default, it is BlobTrigger. For this

function, set the function name to BlobTriggerJs, as shown in Figure 3-18.

Figure 3-18. Naming the function

Click “Create new local app setting,” as shown in Figure 3-19.

Figure 3-19. Creating a new local app setting

Since you have already connected to Azure in the previous section,

you should now directly see all the subscriptions available in Azure. Select
the Azure subscription under which you want to create this function, and
select the Azure storage account with which you want this function to
connect, as shown in Figure 3-20.

44
Chapter 3 Understanding Azure Functions Triggers and Bindings

Figure 3-20. Selecting the Azure Storage account

After selecting the app setting name, the function asks you to provide
the name of the blob that triggers this function. Provide the name of your
blob, as shown in Figure 3-21.

Figure 3-21. Naming your blob

Select “Add to workspace,” as shown in Figure 3-22.

Figure 3-22. Adding to the workspace

Your function is ready. Click the file icon to see all the function files, as
shown in Figure 3-23.

45
Chapter 3 Understanding Azure Functions Triggers and Bindings

Figure 3-23. Function files

To resize the image, you need to add a few extension packages. The
packages that you need to install are @azure/storage-blob, urijs,
stream, jimp, and async. You can install these packages using npm i
<package name>.
Once the packages are installed, copy and paste the following code.

var storage = require('azure-storage');

var URI = require('urijs');
const stream = require('stream');
const Jimp = require('jimp');
var async = require('async');
module.exports = async function (context, myBlob) {
    context.log("JavaScript blob trigger function processed
blob \n Name:", context.bindingData.name, "\n Blob Size:",
myBlob.length, "Bytes");
    var blobService = storage.createBlobService(process.env.
AzureWebJobsStorage);
    var blockBlobName = context.bindingData.name;

46
Chapter 3 Understanding Azure Functions Triggers and Bindings

const widthInPixels = 60;

47
Chapter 3 Understanding Azure Functions Triggers and Bindings

                    });
                });
            }
        ],
        function (err, result) {
            if (err) {
                callback(err, null);
            } else {
                callback(null, result);
            }
        }
    );
};
function generateSasToken(container, blobName, permissions) {
    var connString = process.env.AzureWebJobsStorage;
    var blobService = azure.createBlobService(connString);
    // Create a SAS token that expires in an hour
    // Set start time to five minutes ago to avoid clock skew.
    var startDate = new Date();
    startDate.setMinutes(startDate.getMinutes() - 5);
    var expiryDate = new Date(startDate);
    expiryDate.setMinutes(startDate.getMinutes() + 60);
    permissions = permissions || storage.BlobUtilities.
SharedAccessPermissions.READ;
    var sharedAccessPolicy = {
        AccessPolicy: {
            Permissions: permissions,
            Start: startDate,
            Expiry: expiryDate
        }
    };

48
Chapter 3 Understanding Azure Functions Triggers and Bindings

var sasToken = blobService.generateSharedAccessSignature(

container, blobName, sharedAccessPolicy);
    return {
        token: sasToken,
        uri: blobService.getUrl(container, blobName,
sasToken, true)
    };
}

Once you run the code and upload the image, you should see the
image getting resized.

Running the Example

With these examples, you created two blob-triggered functions running on
the 2.0 framework, one with C# and another with JavaScript/Node.js.
The main thing to note here is that the file host.json is important. It
stores the version of the function and lets the framework know on what
version you are running your function.

{
"version": "2.0"
}

Summary
You should now understand the concept of Azure triggers and bindings,
and you have created a blob-triggered function. The next chapter looks at
creating serverless APIs using Azure Functions.

49
CHAPTER 4

Serverless APIs Using

Azure Functions
Before you start creating APIs with Azure Functions, you must understand
where Azure Functions as a serverless API fit into the current system
architecture you plan to build for your product or applications.
Traditionally, applications were based on a monolithic architecture
because developers wanted all the APIs to be a single deployable unit. Setting
up an individual API for each business scenario was considered a mammoth
task, but the monolithic approach became less desirable with the advent of
cloud computing and agile methodologies. Developers started looking at
microservice architecture because cloud platforms such as Microsoft Azure,
AWS, and GCP made building and managing microservices easy.
In this chapter, we will cover the following topics.

• Monolithic architecture vs. microservice architecture

• Converting monolithic applications to highly scalable

APIs using Azure Functions

• Creating an HTTP-triggered function

• Overview of proxies in Azure Functions

Next, let’s look at monolithic and microservice architectures and

determine which architecture to use in specific circumstances and where
Azure Functions fits in.

© Rahul Sawhney and Kalyan Chanumolu 2023 51

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_4
Chapter 4 Serverless APIs Using Azure Functions

onolithic Architecture vs.

M
Microservice Architecture
The monolithic approach used to be one of the most popular approaches
to building applications. The complete application resides in one codebase
consisting of client-side applications, server-side applications, and
database code.
But with time, these monolithic applications become complex
and difficult to maintain compared to the agile development model.
Monolithic applications are more vulnerable to bugs and deployment
issues. For example, if there is a bug in the client-side code, you still deploy
all the server-side code after fixing the bug since everything resides in
one codebase. This means frequent downtimes for the application and
expensive hardware infrastructure required to ensure high availability.
Also, with most applications moving to the cloud, the monolithic
architecture makes it difficult (and more expensive) for applications to
scale. In addition, DevOps has become slow and complex, and the time
to deploy features, bugs, and hotfixes keep increasing. This is where the
microservice architecture comes to the rescue.
A microservice architecture breaks a complex monolithic application
into small and independent applications. With a microservice architecture,
deploying and scaling individual applications becomes easier and less
expensive and makes DevOps less time-consuming. If you further break
down microservices, they are called nano services.
The following lists the benefits of microservices.

• A small and independent codebase for each

component is easy to maintain.

• Onboarding new developers becomes easy.

• Each service can be scaled individually in the cloud
based on its consumption.

52
Chapter 4 Serverless APIs Using Azure Functions

• Multiple teams can work in parallel on different

microservices.

• Microservices can be written based on the

programming language that best suits the business
scenario. This practice of writing code in multiple
languages to capture additional functionality and
efficiency that is not available in a single language is
known as polyglot programming.

But, with the benefits, there are also trade-offs when using
microservices.

• Writing test cases becomes difficult for each

application.

• Communication within the APIs can become slow and

difficult to troubleshoot if not developed correctly.

• If DevOps is not properly set up, deployment can

become messy and create many issues (but if
done properly, it becomes easy to maintain). An
enterprise application can have more than 10 to 12
microservices, so it is imperative for you to have a
stable CI/CD pipeline for each; otherwise, deploying
these microservices can end up becoming your biggest
blocker.

Figure 4-1 illustrates the differences between monolithic and

microservice architecture.

53
Chapter 4 Serverless APIs Using Azure Functions

Figure 4-1. Monolithic architecture vs. microservice architecture

onverting Monolithic Applications

C
to Highly Scalable APIs Using
Azure Functions
Let’s now look at converting a monolithic e-commerce website to
microservices. A basic e-commerce website comes with a client interface,
a customer profile, product details, checkout, payment functions, and
inventory management.
By looking at these, you can easily see that each component is an
individual business scenario and can be converted to a microservices
architecture. You would have the following microservices.

• Customer service, which includes customer details,

orders, and so on

• Product service

54
Chapter 4 Serverless APIs Using Azure Functions

• Payment and checkout service

• Inventory management service

You can expose each of these services as an API that your front-end
user interface can easily consume.
Now, let’s see how microservices help you scale the application with
a minimum cost. Let’s say you have a holiday season sale coming up, and
you estimate that you have double the amount of traffic on the website
during this time.
After some analysis, you find that the product/service and the payment
and checkout services have the most load, and there won’t be much
change in load on the customer and inventory management service.
With microservices, you can scale out only those two services (product,
payment, and checkout) and leave the other services as is. In contrast,
if you had a monolithic application, you would have to scale out the
complete application, which would greatly increase your operating costs.
Since now you know why you want to convert a monolithic application
to microservices, let’s examine how Azure Functions can help you achieve
that in a simpler and more cost-efficient way.
Azure Functions allows you to write and deploy small pieces of code.
With the help of Azure Functions, you can divide the microservices into
small parts and write a function that performs a specific task. For example,
the customer service microservice would have different activities, such as
Update Profile, View My Orders, Cashback Amount, Card Details, and so
on. You could write each one as a separate function, and on days when you
have a big sale, you can scale up the individual functions.
With Azure Functions, the infrastructure maintenance is taken care of
by the cloud service provider, and you won’t have to worry about scaling
up, upgrading the software, and so on. This reduces the team’s workload
and helps them concentrate on the business.

55
Chapter 4 Serverless APIs Using Azure Functions

Azure Functions provides a free monthly grant of 400,000 GB/s (a unit

of resource consumption) of execution time and one million executions,
which should be enough to run a medium-sized application on Azure
Functions at no cost.
With Azure Functions, each function is completely isolated; so if a bug
or issue is fixed in one function, you do not have to deploy the complete
microservice or application. This is where Azure Functions also makes
DevOps a lot easier.
As shown in Figure 4-2, you can create separate function(s) for each of
the microservices and expose them as REST APIs.

Figure 4-2. Isolated functions

To expose functions as REST APIs, you must create HTTP-triggered

functions so that you can use them. HTTP-triggered functions work like
any other API where you call an endpoint, and it returns the results. Let’s
create an HTTP-triggered function in the next section.

56
Chapter 4 Serverless APIs Using Azure Functions

reating an HTTP-Triggered Function

C
with SQL Server Interaction
Before you start creating the HTTP-triggered function in this section, let’s
first create the Azure SQL Server database with AdventureWorks content so
that you can fetch and modify the data using the HTTP-triggered function.

reating a SQL Server Instance

C
with Sample Data
Let’s get started.
Log in to the Azure portal and click “Create a resource.” Select
Databases from the vertical pane and then select SQL Database, as shown
in Figure 4-3.

Figure 4-3. Selecting SQL Database

57
Chapter 4 Serverless APIs Using Azure Functions

Provide a name for the database. Select a server or create one if it

does not exist. Configure the other settings as shown in Figure 4-4a and
Figure 4-4b.

Figure 4-4a. Creating the database

58
Chapter 4 Serverless APIs Using Azure Functions

Figure 4-4b. Creating the database from the sample

For the data source, you could choose None and create the tables from
scratch. In this case, let’s use the sample database AdventureWorksLT
as shown in Figure 4-4b, because it has tables pre-populated with
sample data.
It takes time for SQL Server and the database to be ready. Once it is
ready, it is under the resource group you selected, as shown in Figure 4-5.

59
Chapter 4 Serverless APIs Using Azure Functions

Figure 4-5. Database created

In the left panel, click “Query editor,” provide your password to

connect, and click OK, as shown in Figure 4-6.

Figure 4-6. Connecting to the database

Once you have successfully logged in, you see the Query Editor. In
the left menu, click Tables. The tables are being created, as shown in
Figure 4-7.

60
Chapter 4 Serverless APIs Using Azure Functions

Figure 4-7. Tables being created

You can query the tables similar to the way it is done in SQL Server
Management Studio.
Your database has been created with some initial data. Let’s now create
an HTTP-triggered function for it.

reating an HTTP-Triggered Function

C
Using C#
Now it’s time to code.
In Chapter 3, you set up your machine to run Azure Functions using
Visual Studio Code. If you did not, please follow the steps in Chapter 3.
Open Visual Studio Code, click the Azure icon, and Create a New
Function as shown in Figure 4-8. Create a folder named HTTP-Triggered-
Function to save your code when prompted.

61
Chapter 4 Serverless APIs Using Azure Functions

Figure 4-8. Creating the function

Select the language to be used for the Azure function. C# is used for
this example, as shown in Figure 4-9.

Figure 4-9. Selecting the language

62
Chapter 4 Serverless APIs Using Azure Functions

Select the .NET 6.0 Isolated LTS runtime version, as shown in

Figure 4-10.

Figure 4-10. Selecting the .Net version

Select the “HTTP trigger” template for the function, as shown in

Figure 4-11.

Figure 4-11. Selecting the HTTP trigger template

Provide a name for the function, as shown in Figure 4-12.

Figure 4-12. Naming the function

63
Chapter 4 Serverless APIs Using Azure Functions

Provide the function’s namespace. By default, it is Company.Function,

but for this function, set it to AzureFunctionBook.Function, as shown in
Figure 4-13.

Figure 4-13. Providing the namespace

Set the access rights for this function. You see three options:
Anonymous, Function, and Admin. The access rights determine which
keys are required to invoke the function.

• Anonymous means no API key is required.

• Function is the default setting if nothing is selected,

which means a function-specific API key is required.

• Admin means a master key is required.

Use Anonymous as the access right for this function, as shown in

Figure 4-14.

Figure 4-14. Setting the access rights

64
Chapter 4 Serverless APIs Using Azure Functions

Let’s add the SqlClient package to the solution. To do this, go to the

terminal and select New Terminal in the top menu. Type dotnet add
package System.Data.SqlClient --version 4.5.1 and press Enter. This
installs the package required for your solution. To verify, go to the .csproj
file. The package was added, as shown in Figure 4-15.

Figure 4-15. Package added

Now everything is set up. You follow a code structure similar to what
you would follow in normal projects. Create two folders: Helper and
Models. In Models, create the CustomerModel.cs file, and in Helper, create
the SqlClientHelper.cs file. You use them for your function, as shown in
Figure 4-16.

65
Chapter 4 Serverless APIs Using Azure Functions

Figure 4-16. Folders created

Click the CustomerModel.cs file and paste the following code.

namespace Company.Function.Models
{
    public class CustomerModel
    {
        public int CustomerID { get; set; }
        public int NameStyle { get; set; }
        public string Title { get; set; }
        public string FirstName { get; set; }
        public string MiddleName { get; set; }

66
Chapter 4 Serverless APIs Using Azure Functions

public string LastName { get; set; }

public string CompanyName { get; set; }
}
}

Click the SqlClientHelper.cs file and paste the following code, which
calls the SQL Server database and gets customer details from the SalesLT.
Customer table.

using System;
using System.Data;
using System.Data.SqlClient;
using Company.Function.Models;
namespace Company.Function.Helper
{
    public static class SqlClientHelper
    {
        public static CustomerModel GetData(int customerId)
        {
            var connection = Environment.GetEnvironmentVariable
("coonectionString");
            CustomerModel customer = new CustomerModel();
            using (SqlConnection conn = new
SqlConnection(connection))
            {
                var text = "SELECT CustomerID, NameStyle,
FirstName, MiddleName, LastName, CompanyName
FROM SalesLT.Customer where CustomerID=" +
customerId;
                SqlCommand cmd = new SqlCommand(text, conn);
                // cmd.Parameters.AddWithValue("@CustomerId",
customerId);

67
Chapter 4 Serverless APIs Using Azure Functions

                conn.Open();
                using (SqlDataReader reader = cmd.
ExecuteReader(CommandBehavior.SingleRow))
                {
                    while (reader.Read() && reader.HasRows)
                    {
                        customer.CustomerID = Convert.
ToInt32(reader["CustomerID"].
ToString());
                        customer.FirstName =
reader["FirstName"].ToString();
                        customer.MiddleName =
reader["MiddleName"].ToString();
                        customer.LastName = reader["LastName"].
ToString();
                        customer.CompanyName =
reader["CompanyName"].ToString();
                    }
                    conn.Close();
                }
            }
            return customer;
        }
    }
}

Go to the main HttpTriggerCSharp.cs file and paste the following

code, which is the function that is triggered when you call it. It first tries
to get the CustomerId value from the query and convert it to int. Then, it
calls the SQLClientHelper.GetData method by passing the CustomerId
value and returning the result.

68
Chapter 4 Serverless APIs Using Azure Functions

using System;
using System.IO;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc;
using Microsoft.Azure.WebJobs;
using Microsoft.Azure.WebJobs.Extensions.Http;
using Microsoft.AspNetCore.Http;
using Microsoft.Extensions.Logging;
using Newtonsoft.Json;
using Company.Function.Helper;
namespace Company.Function
{
    public static class HttpTriggerCSharp
    {
        [FunctionName("HttpTriggerCSharp")]
        public static async Task<IActionResult> Run(
            [HttpTrigger(AuthorizationLevel.Anonymous,
"get", "post", Route = null)] HttpRequest req,
ILogger log)
        {
            log.LogInformation("C# HTTP trigger function
processed a request.");
            int customerId = Convert.ToInt32(req.
Query["customerId"]);
            return (ActionResult) new
OkObjectResult(SqlClientHelper.
GetData(customerId));
        }
    }
}

69
Chapter 4 Serverless APIs Using Azure Functions

If you look at the code, you’ll see you have created a basic customer
profile function where you get customer details based on CustomerID.
Select Debug in the top menu and click Start Debugging. Once the
function is compiled, you see the local URL of the function, as shown in
Figure 4-17.

Figure 4-17. Local URL

Copy this URL, append ?customerId=1 to it, and hit Enter. You should
see the output shown in Figure 4-18.

Figure 4-18. Output

70
Chapter 4 Serverless APIs Using Azure Functions

Now you should understand how to create an HTTP-triggered API

using Azure Functions. In the next section, you look at how you can use
Azure Functions as an OData API to access SQL Server.

reating an HTTP-Triggered OData API

C
for SQL Server Using Azure Functions
Before you start creating functions, you should first understand what
OData is. It is short for Open Data Protocol and defines a set of best
practices for consuming and building web APIs.

Note For more information about OData, you can visit the official
page at https://www.odata.org/.

To create a function, follow the steps in the previous section. The

only change here is that instead of C#, JavaScript is the language for the
function.
Once the function is created, install the following npm packages.

1. Azure-odata-sql
2. Async

3. Tedious

4. Tedious-connection-pool

To install the packages, open the terminal and type the following,
which installs the package for you.

npm install <package name>

71
Chapter 4 Serverless APIs Using Azure Functions

Once the packages are installed, create a file named functions.js and
paste the following code. It connects to the SQL Server database, runs the
query, and returns the data. In the following code, you first create the pool
of SQL connections using poolConfig. Then, you write the getSqlResults
method, which fetches the records from the table.

// TEDIOUS
var ConnectionPool = require('tedious-connection-pool');
var Connection = require('tedious').Connection;
var Request = require('tedious').Request;
var TYPES = require('tedious').TYPES;
// Pool Connection Config
var poolConfig = {
    min: 1,
    max: 10,
    log: true
};
//Connection Config
var config = {
    userName: process.env.databaseUser,
    password: process.env.databasePassword,
    server: process.env.databaseUrl,
    options: {
        database: process.env.databaseName,
        encrypt: true,
        requestTimeout: 0,
    }
};
//create the pool
var pool = new ConnectionPool(poolConfig, config);
pool.on('error', function (err) {
    console.error(err);
});
72
Chapter 4 Serverless APIs Using Azure Functions

function getSqlResult(sqlObject, callback) {

73
Chapter 4 Serverless APIs Using Azure Functions

module.exports = {
getSqlResult: getSqlResult,
}

Go to the main index.js file and paste the following code. In the
following code, you are first configuring the table and schema used in
this OData API. Then you are fetching the pageSize, filters, selection,
ordering, and so on from the query parameters. Once you get all this, you
prepare the query and call azureOdata.format to convert this to a proper
SQL query.

var azureOdata = require('azure-odata-sql');

var async = require('async');
var tableConfig = {
    name: 'Customer',
    schema: 'SalesLT',
    flavor: 'mssql',
};
var defaultPageSize = 30;
module.exports = function(context, req) {
    var module = require('./functions');
    var pageSizeToUse = req.query !== null && req.
query.$pageSize !== null && typeof req.query.$pageSize !==
"undefined" ? req.query.$pageSize : defaultPageSize
    var getSqlResult = module.getSqlResult;
    var query = {
        table: 'Customer',
        filters: req.query !== null && req.query.$filter !==
null && typeof req.query.$filter !== "undefined" ? req.
query.$filter : ",
        inlineCount: "allpages",
        resultLimit: pageSizeToUse,

74
Chapter 4 Serverless APIs Using Azure Functions

skip: req.query !== null && req.query.$page !== null &&

typeof req.query.$page !== "undefined" ? pageSizeToUse
∗ (req.query.$page -1): ",
        take: pageSizeToUse,
        selections: req.query !== null && req.query.$select !==
null && typeof req.query.$select !== "undefined" ? req.
query.$select : ",
        ordering: req.query !== null && req.query.$orderby !==
null && typeof req.query.$orderby !== "undefined" ?
req.query.$orderby : 'CustomerID',
    };
    var statement = azureOdata.format(query, tableConfig);
    var calls = [];
    var data = [];
    async.series([
        function (callback) {
            getSqlResult(statement[0], (err, result) => {
                if (err)
                    throw err;
                data.push(result);
                callback(err, result);
            });
        },
        function (callback) {
            getSqlResult(statement[1], (err, result) => {
                if (err)
                    throw err;
                data.push(result);
                callback(err, result);
            });
        }
    ],

75
Chapter 4 Serverless APIs Using Azure Functions

function (err, result) {

            if (err) {
                console.log(err);
            } else {
                var count = result[0].length;
                context.res = {
                    status: 200,
                    body: {
                    // '@odata.context': req.protocol + '://' +
req.get('host') + '/api/$metadata#Product',
                    'value': result[0],
                    'total': result[1][0].count,
                    'count': count,
                    'page': req.query !== null && req.
query.$page !== null ? req.query.$page : 1
                    },
                    headers: {
                        'Content-Type': 'application/json'
                    }
                };
            }
            context.done();
        });
}

You first prepared the query inside var query = {}. Now, you
convert this into a SQL-understandable query by calling azureOdata.
format(query, tableConfig). Once the query is converted to a SQL
query, you pass this to the function you wrote in functions.js, which runs
the SQL statements and return the data.

76
Chapter 4 Serverless APIs Using Azure Functions

Run Azure Functions by going to the top menu and clicking Debug ➤
Start Debugging. Once the function starts, you make an HTTP call.

http://localhost:7071/products?$filter=CustomerID eq 1&$select=
CustomerID,FirstName,LastName

If you look at this URL, you see two query parameters. One is $filter,
which is converted to a WHERE clause. The other is $select, which is
converted to a SELECT statement.
Once you run the previous query, you get the result shown in
Figure 4-19.

Figure 4-19. Query output

You have created two HTTP-triggered functions: a normal HTTP-

triggered function with C#, and an advanced HTTP-triggered function
using OData with Node.js.

Summary
In this chapter, you learned the differences between monolithic and
microservices architectures. You also created serverless APIs using Azure
Functions. The next chapter looks at the Durable Functions extension and
how you can use durable functions for long-running tasks.

77
CHAPTER 5

Azure Durable
Functions
Over the course of this chapter, we will cover the following durable
function topics.

• Overview

• Bindings

• Performance and scale

• Creating through the Azure portal

• Disaster recovery and geodistribution

Overview of Durable Functions

Durable Functions is an extension to Azure Functions that helps you build
reliable, stateful apps in a serverless environment and define workflows
in your code. The Azure Functions platform manages checkpoints, states,
and restarts for you.
Durable Functions uses a new type of function called an orchestrator
function, which lets you define stateful workflows in code and allows you
to call other functions both synchronously and asynchronously.
The primary use of Durable Functions is to simplify stateful
coordination problems in the serverless world.

© Rahul Sawhney and Kalyan Chanumolu 2023 79

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_5
Chapter 5 Azure Durable Functions

Types of Functions
The Durable Functions extension allows the stateful orchestration
of functions. Each function is made up of a combination of different
functions. Each function plays a different role in orchestration, as shown in
Figure 5-1.

Figure 5-1. The three types of functions

There are three types of functions.

• A client function is the entry point for creating an

instance of a durable function. They are triggered
functions that create a new instance of an orchestration
process. Any available trigger in Azure Functions can
trigger client functions. Also, client functions have
an orchestration binding that allows them to manage
durable orchestrations.

• An orchestrator function describes the order in

which actions are executed. An orchestrator function
must be triggered by an orchestration trigger (a client
function with orchestration binding). Each instance
of an orchestrator has an instance identifier that can
be autogenerated or user-generated and is used to
manage instances of orchestration.
• An activity function is the basic unit of work in durable
function orchestration and are the functions and tasks
that are being orchestrated or ordered in the process.

80
Chapter 5 Azure Durable Functions

For example, you can create a durable function for

order cancellation to handle canceling the shipment,
updating the inventory, and refunding the payment.
Each of these tasks is an activity function, and the
output of one function can be used as the input of
another. An activity trigger must trigger an activity
function.

Application Patterns
The Durable Functions extension caters to five application patterns.

• Function chaining

• Fan-out/fan-in

• Async HTTP APIs

• Monitoring

• Human interaction

Function Chaining
Function chaining is a pattern where you execute functions in sequential
order. Also, you use function chaining when the output of one function
must be used as the input of another function.
Let’s see an example of e-commerce order processing. First, a
customer orders a product, and after that, internally, you process the order
and notify the dealer. Once the dealer confirms that the product is ready to
be shipped, you notify the delivery service to pick up the order and ship it.
Once the product is shipped, you notify the customer. This process can be
done using function chaining, as shown in Figure 5-2.

81
Chapter 5 Azure Durable Functions

Figure 5-2. An example of function chaining

The following code calls the function chaining pattern using C#.

public static async Task<object>

Run(DurableOrchestrationContext context)
{
try
{
var orderProcessedResult = await context.CallActivityAsync
<object>("ProcessOrder");

var dealerNotificationResult = await context.CallActivity

Async<object>("NotifyDealer", orderProcessedResult);

var deliveryServiceResult = await context.Call

ActivityAsync<object>("NotifyDeliveryService",
dealerNotificationResult);

return await context.CallActivityAsync<object>("Notify

Customer", deliveryServiceResult);
}
catch (Exception ex)
{
    // This will be the 5th function which will rollback all
the operations before the function which caused the error
    await context.CallActivityAsync<object>("Rollback", null);
    context.log("Error cannot be processed");
}
}

82
Chapter 5 Azure Durable Functions

Fan-Out/Fan-In
The fan-out/fan-in pattern refers to executing multiple functions in
parallel and then waiting for them to execute. Usually, aggregation work is
done on the result returned by multiple functions.
With normal functions in Azure Functions, fanning out can be done by
publishing multiple messages to the queue. But the fanning is complicated
because you must track when the message is picked up and processed and
store the result. This is a difficult task in Azure Functions, but the Durable
Functions extension handles this pattern quite easily.
Let’s look at an example where you have replenished the stock and
want to notify all the customers who selected “Notify me once the product
is available,” as shown in Figure 5-3.

Figure 5-3. Fan-Out/Fan-In example

This first function updates or replenishes the stock (i.e., products).

Then you call the F2 function for each product that was out of stock and
call multiple functions. One function sends an email, another sends
an SMS message to the customer, and one stacks the product based
on the user interest shown as per the “Notify me once the product is
available” selection. Once you get a response from all three functions,

83
Chapter 5 Azure Durable Functions

you call the UpdateStatus function, which updates the notification status
corresponding to each user who opted for notification, as shown in the
following code.

public static async Task Run(Durableorchestrationcontext ctx)

{
var parallelTasks = new List<Task<int>>();

object[] workBatch = await ctx.CallFunctionAsync<object[]>

("StockUpdate");

for (int i = 0; i < workBatch.Length; i++)

{
Task<int> task = ctx.CallFunctionAsync<int>("F2",
workBatch[i]);
parallelTasks.Add(task);
}
await Task.WhenAll(parallelTasks);

//aggregate result of all tasks and send result to UpdateStatus

int sum = parallelTasks.Sum(t => t.Result);await
ctx.CallFunctionAsync("UpdateStatus", sum);
}

Async HTTP APIs

The Async HTTP APIs pattern takes care of the problem of keeping the
state of long-running processes with external clients. The common way to
implement this pattern is to trigger the long-running job with the HTTP
client and then redirect the external client to another page, which keeps
polling the long-running job’s state.

84
Chapter 5 Azure Durable Functions

The Durable Functions extension provides a built-in capability that

simplifies the code you write for interacting with long-running processes.
Since the Durable Functions runtime manages the state, you don’t have to
implement your own state-tracking mechanism.
Let’s look at an example of a food-ordering app. You order your food,
and the app takes you to a page where you track the order’s status. The
first state is whether the restaurant accepts the order. Once the order
is accepted, it starts showing you the time it takes for the order to be
prepared by the restaurant. Then once the order is ready and picked up by
the delivery person, it shows you a map with the location of the delivery
person. You can implement this with the help of Durable Functions, as
shown in Figure 5-4.

Figure 5-4. Async example

In Figure 5-4, the OrderFood function act as an HTTP API that is called
once the end user clicks Order Food. The OrderFood function checks for
the validity of the order and calls the OrderProcess function. This function
keeps on updating the status of the order.
You redirect the end user to the order-tracking page, which polls the
GetStatus function and shows the order status.

85
Chapter 5 Azure Durable Functions

The following is the demo code depicting the creation of an

orchestrator function for OrderProcess. The following code is part of the
HTTP-triggered OrderFood function.

public static async Task<HttpResponseMessage> Run(

HttpRequestMessage req, DurableOrchestrationClient starter,
ILogger log)
{
// Function name comes from the request URL.
// Function input comes from the request content.
dynamic eventData = await req.Content.ReadAsAsync<object>();
string instanceId = await starter.StartNewAsync("OrderProcess",
eventData);

log.LogInformation($"Started orchestration with

ID = '{instanceId}'.");

return starter.CreateCheckStatusResponse(req, instanceId);

}

Monitoring
The Monitoring pattern is used when you need polling until a condition is
met. A regular timer trigger (which lets you run a function on a specified
schedule) can be used for scenarios such as a cleanup job. But the problem
with this is that the time interval is static, so managing the lifetime of the
instances becomes complex.
On the other hand, the Durable Functions runtime comes with flexible
intervals and lifetime management of tasks. It allows you to create multiple
monitor processes from a single orchestration (see Figure 5-5).

86
Chapter 5 Azure Durable Functions

Figure 5-5. Monitoring example

Human Interaction
Usually, you automate processes that require no human intervention
because people are not as highly available and responsive as cloud
services. But, in certain scenarios that require approval, human
intervention is required, so the automated processes must account for that.
Automated processes generally do this by using timers and
compensation logic. Let’s look at a “leave approval” workflow as an
example. In this case, an employee applies for leave. The notification goes
to the manager to approve it. Here you can have two scenarios. One is if
the manager does not approve it within 48 hours, the leave is automatically
approved. The other scenario is if the manager does not approve it within
48 hours, then it is escalated to the manager’s manager (see Figure 5-6).

87
Chapter 5 Azure Durable Functions

Figure 5-6. An example of a human interaction pattern

The following is the example code.

public static async Task Run(DurableOrchestrationContext

context)
{
await context.CallActivityAsync("SubmitLeaveRequest");

using (var timeoutCts = new CancellationTokenSource())

{
    DateTime dueTime = context.CurrentUtcDateTime.AddHours(48);
    Task durableTimeout = context.CreateTimer(dueTime,
timeoutCts.Token);
    Task<bool> approveLeaveEvent = context.WaitForExternalEvent
<bool>("ApproveLeaveEvent");
    if (approveLeaveEvent == await Task.WhenAny
(approveLeaveEvent, durableTimeout))
    {
        timeoutCts.Cancel();
        await context.CallActivityAsync("ProcessLeaveApproval",
approveLeaveEvent.Result);
    }

88
Chapter 5 Azure Durable Functions

    else
    {
        await context.CallActivityAsync("EscalateEvent");
    }
}
}

Bindings for Durable Functions

The Durable Functions extension introduces two new trigger bindings that
control the execution of orchestrator and activity functions. The Durable
Functions extension also introduces one output binding that triggers the
Durable Functions runtime.

Activity Triggers
An activity trigger enables you to author functions that are called by
orchestrator functions. Activity functions are like any other normal
function. The only difference is that you have ActivityTrigger, which is
triggered from the orchestrator function.
Internally, the following activity trigger binding keeps polling a series
of queues in the default storage account of the function app. The queues
are internal implementations of the extension, so that’s why they are not
part of the orchestrator trigger binding.
The following JSON object in the bindings array defines the activity trigger.

{
  "name": "Input parameter name",
  "activity": "<Optional parameter. Name of the activity",
  "type": "activityTrigger",
  "direction": "in"
}

89
Chapter 5 Azure Durable Functions

The following describes the trigger behavior.

• Threading is an activity trigger like any other function

in Azure Functions that you code and has no limitation
on threading or I/O.

• Message visibility is dequeued and kept invisible for

a configurable amount of time. As long as the function
app is running and is in a healthy state, the visibility of
the messages is renewed automatically.

• Return values are JSON serialized and persisted in the

Azure storage orchestration history table.

The following is the basic code for an activity trigger.

[FunctionName("City_Travel")]
public static string Run([ActivityTrigger] string cityName,
TraceWriter log)
{
log.Info($"I am travelling to {cityName}.");
return $"I am travelling to {cityName}!";
}

Orchestration Triggers
As the name suggests, an orchestration trigger enables you to author
orchestrator functions. The trigger allows you to start new instances of
orchestrator functions and also allows you to resume existing instances
of orchestrator functions that are awaiting a task.
Behind the scenes, the following orchestrator trigger binding keeps
polling a series of queues in the default storage account of the function
app. The queues are internal implementations of the extension, so that’s
why they are not part of the orchestrator trigger binding.

90
Chapter 5 Azure Durable Functions

The orchestrator trigger is defined by the following JSON object in the

bindings array.

{
  "name": "Input parameter name",
  "orchestration": "Optional parameter - Name of
orchestration",
  "type": "orchestrationTrigger",
  "direction": "in"
}

The following describes the trigger behavior.

• Single threading: A single dispatcher thread is used

for all orchestrator functions running on a single host
instance. For this reason, the orchestrator function
code should not perform any I/O. Also, this thread
should not do async work except when awaiting
Durable Functions–specific task types. JavaScript
orchestrator functions should never be declared async.

• Message visibility: The messages are dequeued and

kept invisible for a configurable amount of time.
As long as the function app is running and is in a
healthy state, the visibility of the messages is renewed
automatically. Orchestration triggers do not support
poison message handling.

• Return values: The orchestrator return values are

JSON serialized and are persisted in the Azure storage
orchestration history table.

91
Chapter 5 Azure Durable Functions

The following is the basic code for an orchestrator trigger.

[FunctionName("Orchestrator_City")]
public static async Task<List<string>> Run(
[OrchestrationTrigger] DurableOrchestrationContext context)
{
    var outputs = new List<string>();
    outputs.Add(await context.CallActivityAsync<string>
("City_Travel", "Hyderabad"));
    outputs.Add(await context.CallActivityAsync<string>
("City_Travel", "New York"));
    outputs.Add(await context.CallActivityAsync<string>
("City_Travel", "Delhi"));

    return outputs;
    // returns
    // "I am travelling to Hyderabad"
    // "I am travelling to New York"
    // "I am travelling to Delhi"
}

As you can see, the previous orchestrator function is calling the activity
function City_Travel and passing the city’s name to it. From how it is
written, it looks like the orchestrator function is calling the City_Travel
activity function directly, but it is actually sending a message to a work-
item queue. The activity function City_Travel polls the queue, and as
soon as it receives the message in the queue, it executes the logic.
Once the activity function completes the logic execution, it sends the
response message to the control queue that the orchestrator function is
polling. As the orchestrator function Orchestrator_City receives the
message via OrchestrationTrigger, it shows the response. This is the
behavior of the durable function.

92
Chapter 5 Azure Durable Functions

Once you start the durable function, it creates four control queues and
one work-items queue, as shown in Figure 5-7.

Figure 5-7. Durable function queues

It also creates two Azure storage tables named

DurableFunctionsHubHistory and DurableFunctionsHubInstances.

Orchestration Client
The orchestrator client is responsible for starting/stopping the orchestrator
function. It is also used to query the status, send events, and purge
instances of the history of the orchestrator function.
The orchestrator client binding allows you to write functions in Azure
Functions that interact with orchestrator functions.
The following JSON object in the bindings array defines the
orchestrator client trigger.

{
  "name": "Name of Input Parameter",
  "taskHub": "Optional Parameter. name of the task hub",
  "connectionName": "Optional Parameter. Name of the connection
string in the app settings",
  "type": "orchestrationClient",
  "direction": "in"
}

93
Chapter 5 Azure Durable Functions

The following is the basic code for the orchestration client.

[FunctionName("OrchestrationClient_Start")]
public static async Task<HttpResponseMessage> HttpStart(
[HttpTrigger(AuthorizationLevel.Anonymous, "get",
post")]HttpRequestMessage req,[OrchestrationClient]
DurableOrchestrationClient starter, TraceWriter log)
{
string instanceId = await starter.StartNewAsync
("Orchestrator_City", null);
log.Info($"Running orchestration with
ID = '{instanceId}'.");

return starter.CreateCheckStatusResponse(req, instanceId);

}

Performance and Scaling of Durable Functions

The Durable Functions extension has unique scaling characteristics that
need to be understood to scale and improve performance. To understand
the scaling behavior, you must first understand some of the underlying
details of the Azure storage provider.

History Table
The history table contains the history of events for all the orchestration
instances running within a task hub. The name of the table is in the
format TaskHubNameHistory. The partition key of this table is derived
from the instance ID of the orchestration function. Since the instance
ID is generated randomly, it ensures the optimal distribution of internal
partitions in an Azure storage table. As the orchestrator function instances
run, new rows are added to this table.

94
Chapter 5 Azure Durable Functions

When an orchestration instance runs, first, the appropriate rows of the

history table are loaded into the memory. These historical events are then
replayed in the orchestrator function to return to the previous checkpoint
state. The event sourcing pattern influences this.

Instance Table
This table contains the statuses of all the orchestrations running within
a task hub. The orchestration function instance ID is the partition key
of this table, and the row key is a fixed constant. There is one row per
orchestration function instance.
This table is consistent with the content of the history table. This table
is used by the GetStatusAsync (.NET) API and the getStatus (JavaScript)
API. Also, it is used by the HTTP status query API.
Using a separate table to efficiently satisfy the instance query operation
is influenced by the command and query responsibility segregation (CQRS)
pattern.

Internal Queue Triggers

Activity and orchestrator functions are triggered by the queues in the
task hub of the Azure Functions app. This provides an “at-least-once”
delivery guarantee of messages. There are two types of queues in Durable
Functions.

• The control queue: There are multiple queues in a task

hub. Control queues are more sophisticated than work-
item queues because control queues trigger the stateful
orchestrator functions. Orchestrator messages are load
balanced across the control queue. In a single poll, a
message can dequeue as many as 32 messages, and if
all those messages belong to a single orchestrator, they
are processed as a batch.

95
Chapter 5 Azure Durable Functions

• The work-item queue: Each task hub has one work-

item queue. This queue behaves like a normal queue.
This queue triggers the stateless activity functions by
dequeuing a single message at a time. When a durable
function scales out to multiple VMs, each VM competes
to acquire work from the work-item queue.

Now that you understand the underlying mechanism, let’s look at how
to scale durable functions.

Orchestrator Scale-Out
Stateless functions like activity functions can be scaled out easily by
adding more VMs. But stateful functions like orchestrator functions are
partitioned across one or more queues for them to scale out. By default, a
task hub can have at most 16 partitions; by default, the partition count is
four. The number of control queues is defined in the host.json file for a
function running on the 2.0 runtime, as follows.

{
"extensions": {
      "durableTask": {
            "partitionCount": 2
      }
  }
}

When you scale out the orchestrator function to multiple instances,

each instance acquires a lock on one of the control queues, and this way, it
ensures that each orchestration instance runs on a single host instance at
a time. In the previous example, a task hub has two control queues, so an
orchestration instance can be load balanced across as many as five VMs.
Additional VMs can be added to increase the capacity of activity functions.

96
Chapter 5 Azure Durable Functions

Generally, orchestration functions are intended to be lightweight, so they

should not require more computing power. It is advisable to create no
more than two to five control queues.
Figure 5-8 depicts how Azure Functions behaves in a scaled-
out manner.

Figure 5-8. Azure Functions’ behavior when scaling out

As you can see in Figure 5-8, all instances compete for the work
from the work-item queue, but only two instances at a time can acquire
messages from the control queue, and each instance locks the single
control queue.

Autoscaling
Durable Functions supports autoscaling via the scale controller. The scale
controller monitors the rate of events and decides whether to scale in or
scale out. In Durable Functions, the scale controller monitors the latency
of each queue by issuing a peek command. If the message latencies are
higher than the threshold, the scale controller keeps adding the instances
until it reaches the partition count.

97
Chapter 5 Azure Durable Functions

In work-item queues, the scale controller keeps adding the VM

instances if the message latencies exceed the threshold, irrespective of the
partition count. The maximum number of instances it can add is 200.

Concurrency Throttling
Azure Functions allows you to run multiple functions concurrently within
a single app instance. The concurrency increases the parallel execution
and reduces the number of “cold starts.” But you should also be mindful
that high concurrency results in high per-VM memory usage.
Orchestrator and activity functions support concurrency, and their
limits can be set in host.json. The setting for an activity function is
maxConcurrentActivityFunctions and for an orchestrator function is
maxConcurrentOrchestratorFunctions.

{
    "extensions": {
        "durableTask": {
            "maxConcurrentActivityFunctions": 20,
            "maxConcurrentOrchestratorFunctions": 20
        }
    }
}

By default, the number of activity and orchestrator function executions

is capped at ten times the number of cores on the VM.

Orchestrator Function Replay

As you know, orchestrator functions are stateful functions, and they replay
to the checkpoint using the contents of the history table. The orchestrator
function code is replayed every time a batch of messages is dequeued from
the control queue by default.

98
Chapter 5 Azure Durable Functions

Durable Functions provides an ability to decrease the aggressive

behavior of the replay by using extended sessions. When you enable
extended sessions, the function instances are held in memory
for that time, and you can process messages without a full replay.
Enabling extended sessions reduces the I/O against the Azure storage
table and thus increases the throughput. You can enable extended
sessions by setting extendedSessionsEnabled to true. To control
how long you keep the idle session in the memory, you use the
extendedSessionIdleTimeoutInSeconds setting in host.json, as follows.

{
    "extensions": {
        "durableTask": {
            "extendedSessionsEnabled": true,
            "extendedSessionIdleTimeoutInSeconds": 30
        }
    }
}

But there are always two sides of a coin. So, when enabling extended
session to increase throughput, there is also a downside.
• It can increase function app memory usage.

• It can decrease throughput if there are many

concurrent, short-lived orchestrator functions.

Performance Targets
If you plan to use durable functions in a production application, consider
the performance requirements early in the process because they define the
pattern you should use for your functions.

99
Chapter 5 Azure Durable Functions

Table 5-1 shows the maximum throughput for various scenarios.

Table 5-1. Maximum Throughput

Scenario Maximum Throughput

Sequential activity execution 5 activities per second per instance

Parallel activity execution (fan-out) 100 activities per second per instance
Parallel response processing (fan-in) 150 responses per second per instance
External event processing 50 events per second per instance

reating Durable Functions Using

C
the Azure Portal
Now that you understand what a durable function is, let’s create one.

Creating a Durable Function

Open the Azure portal and click “Create a resource.” Next, select Compute
➤ Function App, as shown in Figure 5-9.

100
Chapter 5 Azure Durable Functions

Figure 5-9. Starting a durable function

101
Chapter 5 Azure Durable Functions

Provide the details shown in Figure 5-10 and click Create.

Figure 5-10. App details

102
Chapter 5 Azure Durable Functions

Once the deployment succeeds, go to the resource and select the

durable-func-new-book function, as shown in Figure 5-11.

Figure 5-11. Selecting the function

Expand the function’s app and click the + icon. Then, click the
“In-portal” environment and continue, as shown in Figure 5-12.

Figure 5-12. Select the environment and continue

103
Chapter 5 Azure Durable Functions

Click “More templates” and click “Finish and view templates and
continue, as shown in Figure 5-12,” as shown in Figure 5-13.

Figure 5-13. Choosing more templates

In the search field, type durable and select the “Durable Functions
HTTP starter” template, as shown in Figure 5-14.

Figure 5-14. Selecting the starter template

104
Chapter 5 Azure Durable Functions

Click Install to install the Durable Functions extension, as shown in

Figure 5-15.

Figure 5-15. Starting the installation

Name the orchestrator client function OrchestrationClient_Start.

Paste the following code and click Save.

#r "Microsoft.Azure.WebJobs.Extensions.DurableTask"
#r "Microsoft.Azure.WebJobs.Extensions.Http"
#r "Newtonsoft.Json"
using System.Net.Http;
using System.Threading.Tasks;
using Microsoft.Azure.WebJobs;
using Microsoft.Azure.WebJobs.Extensions.Http;
using Microsoft.Azure.WebJobs.Host;
using Microsoft.Extensions.Logging;

[FunctionName("OrchestrationClient_Start")]
public static async Task<HttpResponseMessage> Run(
[HttpTrigger(AuthorizationLevel.Anonymous, "get",
"post")] HttpRequestMessage req,
[OrchestrationClient] DurableOrchestrationClient
starter, ILogger log)
105
Chapter 5 Azure Durable Functions

{
    string instanceId = await starter.StartNewAsync
("Orchestrator_City", null);
    log.LogInformation($"Running orchestration with
ID = '{instanceId}'.");
    return starter.CreateCheckStatusResponse(req, instanceId);
}

Click the + icon again and type durable. Select “Durable Functions
orchestrator,” as shown in Figure 5-16.

Figure 5-16. Selecting the orchestrator

Name the function Orchestrator_City and paste the following code.

#r "Microsoft.Azure.WebJobs.Extensions.DurableTask"
using System.Collections.Generic;
using System.Threading.Tasks;

[FunctionName("Orchestrator_City")]
public static async Task<List<string>> Run(

106
Chapter 5 Azure Durable Functions

[OrchestrationTrigger] DurableOrchestrationContext context)

{
    var outputs = new List<string>();
    outputs.Add(await context.CallActivityAsync<string>
("City_Travel", "Hyderabad"));
    outputs.Add(await context.CallActivityAsync<string>
("City_Travel", "New York"));
    outputs.Add(await context.CallActivityAsync<string>
("City_Travel", "Delhi"));

    return outputs;
    // returns
    // "I am travelling to Hyderabad"
    // "I am travelling to New York"
    // "I am travelling to Delhi"
}
Click the + icon again and type durable. Select “Durable Functions
activity,” as shown in Figure 5-17.

Figure 5-17. Selecting the activity

107
Chapter 5 Azure Durable Functions

Name the function City_Travel and paste the following code.

#r "Microsoft.Azure.WebJobs.Extensions.DurableTask"
[FunctionName("City_Travel")]
public static string Run([ActivityTrigger] string cityName,
ILogger log)
{
log.LogInformation($"I am travelling to {cityName}.");
return $"I am travelling to {cityName}!";
}

Click the function name and then click “Platform features.” In the
Platform features tab, select App Service Editor, as shown in Figure 5-18.

Figure 5-18. Selecting App Service Editor

108
Chapter 5 Azure Durable Functions

The App Service Editor opens in a new tab. Now, select the host.json
file and copy and paste the following code into it.

{
    "version": "2.0",
    "logging": {
        "fileLoggingMode": "always",
        "logLevel": {
          "default": "Information",
          "Host.Results": "Information",
          "Function": "Information",
          "Host.Aggregator": "Trace"
        }
      }
}

There’s no need to save the file. It autosaves, as shown in Figure 5-19.

Figure 5-19. The code

Next, return to the function app and go to the OrchestrationClient_

Start function. Click Get Function Url and copy the URL. The URL is in
the format of https://<function-http-instance>/api/orchestrators/
{functionName}?code=#code. Replace {functionName} with the name

109
Chapter 5 Azure Durable Functions

of the orchestration HTTP trigger, which is OrchestrationClient_Start.

Paste the URL in the browser and press Enter to execute a new orchestrator
function instance.

isaster Recovery and Geodistribution

D
of Durable Functions
Since you have deployed your Azure durable function and want to
use it in production, you should now look at how you can make it
production-ready.
Whenever you want a solution to run on a cloud service provider such
as Microsoft, Amazon, Google, and so on, you should specifically plan
for disaster recovery and make sure your application is running in case
there is any disaster and the region in which application is running in
goes down.
Also, make sure that the data to be stored to make this application run
successfully is properly georeplicated.
To enable disaster recovery for durable functions, you should first
ensure that your function is stateless. Once you have done that, you can
enable disaster recovery by leveraging another Microsoft service called
Traffic Manager.
You can configure Azure Functions as an app service in Traffic
Manager and use any routing strategy.
For geodistribution, you should always consider keeping a copy of the
data in multiple regions. All the Azure data storage services, such as Azure
Storage, Azure Cosmos DB, and Azure SQL, provide georeplication of data
across Azure data centers. You can enable these georeplication services to
ensure that your data is available in multiple regions. This helps you make
your function available quickly in a disaster.

110
Chapter 5 Azure Durable Functions

Summary
This chapter covered how durable functions work and the patterns you’ll
use with the Durable Functions extension. Also, in this chapter, you
created your first durable function running in Azure. You now understand
how to manage your functions in the event of a disaster.
The next chapter looks at deploying functions to Azure using a CI/CD
pipeline.

111
CHAPTER 6

Deploying Functions
to Azure
This chapter covers the following topics.

• Deploying functions to Azure using continuous

deployment

• Deploying functions to Azure using ARM templates

This chapter walks you through the ways to deploy functions to Azure.
By the end of this chapter, you should be able to deploy your functions in
two different ways.

eploying Functions Using

D
Continuous Deployment
Azure Functions Continuous integration/continuous deployment (CI/CD)
integrates seamlessly with continuous integration/continuous deployment
(CI/CD) and the Azure pipeline, which allows you to continuously deploy
your functions to production. Continuous deployment makes it easier to
deploy code bits in a project where multiple people are working on the
same codebase, and changes to the code repository are frequent.

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_6
Chapter 6 Deploying Functions to Azure

Using App Service continuous integration, you can easily deploy a

function app. Azure Functions integrates seamlessly with the following
deployment sources.

• Azure DevOps (a.k.a. VSTS)

• OneDrive

• GitHub

• Dropbox

• Bitbucket

• Git local repository

• External repositories such as Mercurial and Git

Continuous deployment is configured on a per–function app basis.

Once the continuous deployment is enabled, the access. to the function
app is set to read-only in the Azure portal.

etting Up a Code Repository

S
for Continuous Deployment
Before you set up continuous deployment for your function app, you
should arrange your source code properly. The name of the directory is the
name of the function app. The host.json file resides in the parent or top
folder. Each subfolder in the function app consists of separate functions. A
bin folder contains library files and packages the function requires to run,
as shown in Figure 6-1.

114
Chapter 6 Deploying Functions to Azure

Figure 6-1. Project organization

All the functions in the function app should have the same
language worker.
Now that your code repository is ready, let’s set up your function for
continuous deployment.

Setting Up an Azure DevOps Account

Before you can set up continuous deployment for your Azure function,
you need to set up your Azure DevOps account to connect it to the Azure
Functions service.
Set up your Azure DevOps account by following these steps.
Go to the Azure portal (https://ms.portal.azure.com/) and click
DevOps. Select “Azure DevOps organizations,” as shown in Figure 6-2a and
Figure 6-2b.

115
Chapter 6 Deploying Functions to Azure

Figure 6-2a. Finding Azure DevOps organizations resource

If you do not already have an Azure DevOps organization, this page

presents you with options to create one.

116
Chapter 6 Deploying Functions to Azure

Figure 6-2b. List your Azure DevOps organizations

Once created, you should see the list of organizations, as shown in

Figure 6-3.

Note Skip the next steps if the AzureDevOps organization does not

show up here after creating one. This can happen if you are using a
free Azure subscription. You can still set up continuous integration.
The steps are covered in the next section.

117
Chapter 6 Deploying Functions to Azure

Figure 6-3. Organizations list

Once you click the organization, a blade should open on the right side,
as shown in Figure 6-4. Select “Set up billing” in the menu.

Figure 6-4. Setting up billing

A vertical blade opens with the available Azure subscriptions. Select a

suitable one and click Link, as shown in Figure 6-5.

118
Chapter 6 Deploying Functions to Azure

Figure 6-5. Linking to an Azure subscription

Once you get a notification that the subscription is linked, you are
good to go, as shown in Figure 6-6.

119
Chapter 6 Deploying Functions to Azure

Figure 6-6. Success

Enabling Deployment Slots

Azure allows you to create deployment slots for Azure Functions and web
apps. Think of a deployment slot as a clone of the production environment
where you can deploy your app first, and test it before deploying it to
the production slot. And if everything looks good, you can just swap the
staging slot with the production slot without downtime. This helps you in
two ways.

• The app can be thoroughly tested before being released

to production.

• If the new deployment after the slot swap misbehaves

in production or has a bug, you have your last working
code already available in the staging slot, so you can
swap it back to production.

Enable deployment slots on your Azure function app by following

these next steps.
On the Azure portal, navigate to your Azure function, select
Deployment Slots under the deployment section, as shown in Figure 6-7,
and click Add Slot. The production slot corresponds to the function app.

120
Chapter 6 Deploying Functions to Azure

Figure 6-7. Deployment slot

Specify a name for the slot and click Add, as shown in Figure 6-8.

Figure 6-8. Add deployment slot

Once the slot is created, navigate to the resource group where your
Azure function is deployed. You should see an identical App Service (Slot)
and the original function app, as shown in Figure 6-9.

121
Chapter 6 Deploying Functions to Azure

Figure 6-9. App Service (Slot)

This new slot is now practically identical to the production app. You
can create up to five deployment slots to test various functionalities and
configurations before deploying to the production slot.

etting up Continuous Deployment

S
for Azure Functions
You have linked your Azure DevOps organization with an Azure
subscription, so you can now set up continuous development for Azure
Functions.
Go to the Azure portal, navigate to the resource group where
the function app is created, select the preproduction slot, and select
Deployment Center, as shown in Figure 6-10.

122
Chapter 6 Deploying Functions to Azure

Figure 6-10. Deployment Center

Now you should see many options, such as Azure Repos, GitHub,
Bitbucket, and so on, from where you can set up CI/CD. For this example,
select Azure Repos and click Next.
Azure now presents you with a choice of selecting a build provider.
Azure App Service build service should be good for basic scenarios, but
Azure Pipelines lets you customize the build artifacts accordingly if you
need more control over the build process.

123
Chapter 6 Deploying Functions to Azure

Figure 6-11. Azure build providers

Select Azure DevOps Organization, Project, Repository, and Branch

menus, and click the Continue button, as shown in Figure 6-12.

124
Chapter 6 Deploying Functions to Azure

Figure 6-12. Adding the Azure DevOps configuration

Check the summary and click Finish, and your continuous deployment
is now ready and set up, as shown in Figure 6-13.

Figure 6-13. CD now ready

Click Build Pipeline.

125
Chapter 6 Deploying Functions to Azure

You are taken to the build pipeline of your function, as shown in

Figure 6-14.

Figure 6-14. Build pipelines

To check the release pipeline, click the Release Pipeline link, as shown
in Figure 6-14. You are taken to the release pipeline of the function in
Visual Studio Team Service (VSTS), as shown in Figure 6-15.

Figure 6-15. Release pipeline

Azure has made configuring CI/CD intuitive and easy to set up with
just a few clicks.

126
Chapter 6 Deploying Functions to Azure

eploying Azure Functions Using

D
ARM Templates
One of the most popular ways of deploying anything on Azure has
been Azure Resource Manager (ARM) templates. Functions can also be
deployed using ARM templates. In this section, you look at the required
parameters and resources that enable you to deploy functions with ARM
templates.
You need the following resources to start deploying functions using
ARM templates.

• An Azure storage account

• A hosting plan

• A function app

Let’s set up everything using ARM templates.

For any deployment on Azure, you require an Azure storage account,
which is the case here. ARM templates first copy the zip file to Azure Blob
Storage and then use that zip file to deploy the required resource.
The following code snippet creates an Azure storage account using an
ARM template.

{
      "type": "Microsoft.Storage/storageAccounts",
      "name": "[variables('storageAccountName')]",
      "apiVersion": "2016-12-01",
      "location": "[parameters('location')]",
      "kind": "Storage",
      "sku": {
        "name": "[parameters('storageAccountType')]"
      }
}

127
Chapter 6 Deploying Functions to Azure

This code is looking for the storageAccountType parameter, which can

be set up in the parameters section in the ARM template, as shown next.

"storageAccountType": {
      "type": "string",
      "defaultValue": "Standard_LRS",
      "allowedValues": ["Standard_LRS", "Standard_GRS",
"Standard_RAGRS"],
      "metadata": {
        "description": "Storage Account type"
      }
}

The Azure storage account is set up, so let’s look at setting up the
hosting plan. Here you have two hosting plans: the Consumption plan and
the App Service plan. Let’s first look at the Consumption plan.

eploying a Function App on the

D
Consumption Plan
The Consumption plan allows you to make the best use of Azure
Functions. It dynamically allocates compute power when your code is
running. It scales out to handle the extra load and then returns to normal
when it lessens. So, if Azure Functions is not running, you are not paying
anything for idle virtual machines. Also, you don’t have to worry about
peak load in advance because the Consumption plan takes care of it.
The Consumption plan is a special type of serverfarm resource,
and in ARM templates, you specify it by setting the Dynamic value for the
computeMode and sku properties.

128
Chapter 6 Deploying Functions to Azure

{
      "type": "Microsoft.Web/serverfarms",
      "apiVersion": "2015-04-01",
      "name": "[variables('hostingPlanName')]",
      "location": "[parameters('location')]",
      "properties": {
        "name": "[variables('hostingPlanName')]",
        "computeMode": "Dynamic",
        "sku": "Dynamic"
      }
}

In addition, the Consumption plan requires two more settings:

WEBSITE_CONTENTAZUREFILECONNECTIONSTRING and WEBSITE_
CONTENTSHARE. These properties configure the storage account and file
path where the function app and configuration are stored.

"properties": {
        "serverFarmId": "[resourceId('Microsoft.Web/
serverfarms', variables('hostingPlanName'))]",
        "siteConfig": {
          "appSettings": [
            {
              "name": "WEBSITE_CONTENTAZUREFILE
CONNECTIONSTRING",
              "value": "[concat('DefaultEndpointsProtocol=
https;AccountName=', variables('storageAccount
Name'), ';AccountKey=', listKeys(variables
('storageAccountid'),'2015-05-01-preview').key1)]"
            },
            {
              "name": "WEBSITE_CONTENTSHARE",

129
Chapter 6 Deploying Functions to Azure

"value": "[toLower(variables('functionAppName'))]"
            }
          ]
        }
}

The complete ARM template to deploy Azure Functions on the

Consumption plan is as follows.

{
  "$schema": "https://schema.management.azure.com/schemas/
2015-01-01/deploymentTemplate.json#",
  "contentVersion": "1.0.0.0",
  "parameters": {
    "appName": {
      "type": "string",
      "metadata": {
        "description": "Function App Name"
      }
    },
    "storageAccountType": {
      "type": "string",
      "defaultValue": "Standard_LRS",
      "allowedValues": ["Standard_LRS", "Standard_GRS",
"Standard_RAGRS"],
      "metadata": {
        "description": "Storage Account type"
      }
    },
    "location": {
      "type": "string",
      "defaultValue": "[resourceGroup().location]",

130
Chapter 6 Deploying Functions to Azure

      "metadata": {
        "description": "Location for all resources."
      }
    },
    "runtime": {
      "type": "string",
      "defaultValue": "node",
      "allowedValues": ["node", "dotnet", "java"],
      "metadata": {
        "description": "The language worker runtime to load in
the function app."
      }
    }
  },
  "variables": {
    "functionAppName": "[parameters('appName')]",
    "hostingPlanName": "[parameters('appName')]",
    "applicationInsightsName": "[parameters('appName')]",
    "storageAccountName": "[concat(uniquestring(resourceGro
up().id), 'azfunctions')]",
    "storageAccountid": "[concat(resourceGroup().id,'/
providers/','Microsoft.Storage/storageAccounts/', variables
('storageAccountName'))]",
    "functionWorkerRuntime": "[parameters('runtime')]"
  },
  "resources": [
    {
      "type": "Microsoft.Storage/storageAccounts",
      "name": "[variables('storageAccountName')]",
      "apiVersion": "2016-12-01",
      "location": "[parameters('location')]",

131
Chapter 6 Deploying Functions to Azure

      "kind": "Storage",
      "sku": {
        "name": "[parameters('storageAccountType')]"
      }
    },
    {
      "type": "Microsoft.Web/serverfarms",
      "apiVersion": "2015-04-01",
      "name": "[variables('hostingPlanName')]",
      "location": "[parameters('location')]",
      "properties": {
        "name": "[variables('hostingPlanName')]",
        "computeMode": "Dynamic",
        "sku": "Dynamic"
      }
    },
    {
      "apiVersion": "2015-08-01",
      "type": "Microsoft.Web/sites",
      "name": "[variables('functionAppName')]",
      "location": "[parameters('location')]",
      "kind": "functionapp",
      "dependsOn": [
        "[resourceId('Microsoft.Web/serverfarms', variables
('hostingPlanName'))]",
        "[resourceId('Microsoft.Storage/storageAccounts',
variables('storageAccountName'))]"
      ],
      "properties": {
        "serverFarmId": "[resourceId('Microsoft.Web/
serverfarms', variables('hostingPlanName'))]",

132
Chapter 6 Deploying Functions to Azure

        "siteConfig": {
          "appSettings": [
            {
              "name": "AzureWebJobsDashboard",
              "value": "[concat('DefaultEndpointsProtocol=
https;AccountName=', variables('storageAccount
Name'), ';AccountKey=', listKeys(variables
('storageAccountid'),'2015-05-01-
preview').key1)]"
            },
            {
              "name": "AzureWebJobsStorage",
              "value": "[concat('DefaultEndpointsProtocol=
https;AccountName=', variables('storageAccount
Name'), ';AccountKey=', listKeys(variables(
'storageAccountid'),'2015-05-01-preview').key1)]"
            },
            {
              "name": "WEBSITE_CONTENTAZUREFILE
CONNECTIONSTRING",
              "value": "[concat('DefaultEndpointsProtocol=
https;AccountName=', variables('storageAccount
Name'), ';AccountKey=', listKeys(variables
('storageAccountid'),'2015-05-01-
preview').key1)]"
            },
            {
              "name": "WEBSITE_CONTENTSHARE",
              "value": "[toLower(variables('functionAppName'))]"
            },

133
Chapter 6 Deploying Functions to Azure

            {
              "name": "FUNCTIONS_EXTENSION_VERSION",
              "value": "~2"
            },
            {
              "name": "WEBSITE_NODE_DEFAULT_VERSION",
              "value": "8.11.1"
            },
            {
              "name": "APPINSIGHTS_INSTRUMENTATIONKEY",
              "value": "[reference(resourceId('microsoft.
insights/components/', variables('application
InsightsName')), '2015-05-01').
InstrumentationKey]"
            },
            {
              "name": "FUNCTIONS_WORKER_RUNTIME",
              "value": "[variables('functionWorkerRuntime')]"
            }
          ]
        }
      }
    },
    {
      "apiVersion": "2018-05-01-preview",
      "name": "[variables('applicationInsightsName')]",
      "type": "microsoft.insights/components",
      "location": "East US",
      "tags": {

134
Chapter 6 Deploying Functions to Azure

"[concat('hidden-link:', resourceGroup().id, '/

providers/Microsoft.Web/sites/', variables('application
InsightsName'))]": "Resource"
      },
      "properties": {
        "ApplicationId": "[variables('applicationInsights
Name')]",
        "Request_Source": "IbizaWebAppExtensionCreate"
      }
    }
  ]
}

eploying a Function App on the App

D
Service Plan
With this plan, Azure Function runs on dedicated VMs similar to web apps.
You can set up the App Service plan in an ARM template, as follows.

{
      "type": "Microsoft.Web/serverfarms",
      "apiVersion": "2016-09-01",
      "name": "[variables('hostingPlanName')]",
      "location": "[parameters('location')]",
      "properties": {
        "name": "[variables('hostingPlanName')]",
        "sku": "[parameters('sku')]",
        "workerSize": "[parameters('workerSize')]",
        "hostingEnvironment": "",
        "numberOfWorkers": 1
      }
}

135
Chapter 6 Deploying Functions to Azure

Here, workerSize is the size of the VM, which is small (0), medium
(1), or large (2). You can set up the worker size in the ARM template in the
parameters section, as follows.

"workerSize": {
      "type": "string",
      "allowedValues": [
        "0",
        "1",
        "2"
      ],
      "defaultValue": "0",
      "metadata": {
        "description": "The instance size of the hosting plan"
      }
}

The following is the complete ARM template for Azure Functions.

136
Chapter 6 Deploying Functions to Azure

        "Free",
        "Shared",
        "Basic",
        "Standard"
      ],
      "defaultValue": "Standard",
      "metadata": {
        "description": "The pricing tier for the hosting plan."
      }
    },
    "workerSize": {
      "type": "string",
      "allowedValues": [
        "0",
        "1",
        "2"
      ],
      "defaultValue": "0",
      "metadata": {
        "description": "The instance size of the hosting plan"
      }
    },
    "storageAccountType": {
      "type": "string",
      "defaultValue": "Standard_LRS",
      "allowedValues": [
        "Standard_LRS",
        "Standard_GRS",
        "Standard_RAGRS"
      ],
      "metadata": {

137
Chapter 6 Deploying Functions to Azure

"description": "Storage Account type"

      }
    },
    "location": {
      "type": "string",
      "defaultValue": "[resourceGroup().location]",
      "metadata": {
        "description": "Location for all resources."
      }
    }
  },
  "variables": {
    "functionAppName": "[parameters('appName')]",
    "hostingPlanName": "[parameters('appName')]",
    "storageAccountName": "[concat(uniquestring(resource
Group().id), 'functions')]"
  },
  "resources": [
    {
      "type": "Microsoft.Storage/storageAccounts",
      "name": "[variables('storageAccountName')]",
      "apiVersion": "2018-02-01",
      "location": "[parameters('location')]",
      "kind": "Storage",
      "sku": {
        "name": "[parameters('storageAccountType')]"
      }
    },
    {
      "type": "Microsoft.Web/serverfarms",
      "apiVersion": "2016-09-01",
      "name": "[variables('hostingPlanName')]",

138
Chapter 6 Deploying Functions to Azure

      "location": "[parameters('location')]",
      "properties": {
        "name": "[variables('hostingPlanName')]",
        "sku": "[parameters('sku')]",
        "workerSize": "[parameters('workerSize')]",
        "hostingEnvironment": "",
        "numberOfWorkers": 1
      }
    },
    {
      "apiVersion": "2016-08-01",
      "type": "Microsoft.Web/sites",
      "name": "[variables('functionAppName')]",
      "location": "[parameters('location')]",
      "kind": "functionapp",
      "properties": {
        "name": "[variables('functionAppName')]",
        "serverFarmId": "[resourceId('Microsoft.Web/
serverfarms', variables('hostingPlanName'))]",
        "hostingEnvironment": "",
        "clientAffinityEnabled": false,
        "siteConfig": {
          "alwaysOn": true
        }
      },
      "dependsOn": [
        "[resourceId('Microsoft.Web/serverfarms', variables
('hostingPlanName'))]",
        "[resourceId('Microsoft.Storage/storageAccounts',
variables('storageAccountName'))]"
      ],

139
Chapter 6 Deploying Functions to Azure

      "resources": [
        {
          "apiVersion": "2016-08-01",
          "name": "appsettings",
          "type": "config",
          "dependsOn": [
            "[resourceId('Microsoft.Web/sites', variables
('functionAppName'))]",
            "[resourceId('Microsoft.Storage/storageAccounts',
variables('storageAccountName'))]"
          ],
          "properties": {
            "AzureWebJobsStorage": "[concat('DefaultEndpoints
Protocol=https;AccountName=',variables('storage
AccountName'),';AccountKey=',listkeys(resourceId
('Microsoft.Storage/storageAccounts', variables
('storageAccountName')), '2015-05-01-preview').
key1,';')]",
            "AzureWebJobsDashboard": "[concat('DefaultEndpoints
Protocol=https;AccountName=',variables('storage
AccountName'),';AccountKey=',listkeys(resourceId
('Microsoft.Storage/storageAccounts', variables
('storageAccountName')), '2015-05-01-preview').
key1,';')]",
            "FUNCTIONS_EXTENSION_VERSION": "~1"
          }
        }
      ]
    }
  ]
}

140
Chapter 6 Deploying Functions to Azure

Once the function is deployed using the CI/CD pipeline and you have
set up the staging slot, the function is deployed to the staging slot. The
following steps take you to the staging slot.
Go to the Azure portal and click Function Apps in the menu, as shown
in Figure 6-16.

Figure 6-16. Selecting Function Apps

Select the function for which you created the CI/CD pipeline. A pipeline
was created for durable-func-new-book, as shown in Figure 6-17.

Figure 6-17. Pipeline

141
Chapter 6 Deploying Functions to Azure

Click “Platform features” and select Deployment Center, as shown in

Figure 6-18.

Figure 6-18. Selecting Deployment Center

In the Deployment Center, click the slot, as shown in Figure 6-19.

Figure 6-19. Selecting the slot

In the slot, you see the URL of the staging slot and the Swap option, as
shown in Figure 6-20. Now, you can test your function in the staging slot.
Once you are satisfied that things are running fine, you can swap the slot.

142
Chapter 6 Deploying Functions to Azure

Figure 6-20. Testing

To swap the slot, click the Swap button, as highlighted in Figure 6-21.
When you click Swap, the vertical screen open with option to swap. Once
you are satisfied with the values, click Swap again.

Figure 6-21. Clicking Swap

You can get the Azure Quick Start template at https://azure.

microsoft.com/en-us/resources/templates/ or from GitHub at
https://github.com/Azure/azure-quickstart-templates/.

143
Chapter 6 Deploying Functions to Azure

Summary
This chapter examined setting up a CI/CD for your Azure function using
Azure DevOps, enabling deployment slots for testing your functions before
making changes to your functions in production. You also learned about
deploying functions using ARM templates. In the next chapter, you look at
what’s required to make functions production-ready.

144
CHAPTER 7

Getting Functions
Production-Ready
This chapter explains logging, Application Insights, securing functions,
and using cross-origin site scripting, and by the end, your function will be
production-ready.
This chapter covers the following topics.

• Using built-in logging

• Using Application Insights to monitor functions

• Securing functions

• Configuring CORS in Azure Functions

Using Built-in Logging

The first thing that comes to mind about monitoring functions is error
logging. You’ll want to log errors in Azure Functions so that you know what
went wrong and can fix it.
By default, Azure Functions has a logger instance that logs errors to
Azure File Storage. The logger is passed to the function along with the
invocation, as shown in Figure 7-1.

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_7
Chapter 7 Getting Functions Production-Ready

Figure 7-1. The logger is passed to the function

As you can see in Figure 7-1, an instance of ILogger is passed as an

argument to the function invocation. You can use the extension methods
from Microsoft.Extensions.Logging to log events. The methods exposed
are LogTrace, LogDebug, LogInformation, LogWarning, LogError, and
LogCritical.
For a JavaScript function, it looks like Figure 7-2.

Figure 7-2. The JavaScript function

The context passed in Figure 7-2 has a log function, which you can use
to log at different levels. The log function has similar levels, such as Trace,
Debug, Information, Warning, Error, Critical, and None.
The host and function logs are stored in a file share in the storage
account associated with the Azure function under /LogFiles/
Application/Functions.

146
Chapter 7 Getting Functions Production-Ready

sing Application Insights to Monitor

U
Azure Functions
Application Insights is an extension of Azure Monitor that provides
capabilities to monitor your application’s performance, availability, and
usage. Azure Functions offers seamless integration with Application
Insights for logging details such as incoming HTTP requests, response
times, failure rates, exceptions, and more.

pplication Insights Settings

A
for Azure Functions
To connect, Azure Functions needs to know the Application Insights
connection string key. The APPLICATIONINSIGHTS_CONNECTION_STRING key
must be set in the Azure Functions app settings.
You can integrate Application Insights with Azure Functions in
two ways.

• Automatic integration while creating a function app

• Manual connection to an existing Application

Insights service

Integrate Application Insights During New

Azure Function Creation
Let’s see how this is done.
Go to the Azure portal and click “Create a resource.” Then, click
Compute ➤ Function App, as shown in Figure 7-3.

147
Chapter 7 Getting Functions Production-Ready

Figure 7-3. Starting the function app

In the Basics tab, provide details about the function app, as shown in
Figure 7-4. (This was covered in Chapter 2.)

148
Chapter 7 Getting Functions Production-Ready

Figure 7-4. Creating an Azure function

In the Monitoring tab, select Enable Application Insights. Next, select

either “Create new” or “Select existing resource” and set up Application
Insights, as shown in Figure 7-5.

149
Chapter 7 Getting Functions Production-Ready

Figure 7-5. Setting up Application Insights

Your new function is now integrated with Application Insights.

anually Connecting Application Insights

M
to Azure Functions
If you already have an Azure function created without Application Insights,
you can manually connect them. Let’s look at that process now.
Go to the Azure portal, click “Create a resource”, and search for
Application Insights. Then click Create, as shown in Figure 7-6.

150
Chapter 7 Getting Functions Production-Ready

Figure 7-6. Creating Application Insights

Once you are on the Application Insights creation page, provide details
such as the name, application type, resource group, and location, as shown
in Figure 7-7. Click Create.

151
Chapter 7 Getting Functions Production-Ready

Figure 7-7. Application Insights properties

Once Application Insights is ready, navigate to the resource and copy

the connection string, as shown in Figure 7-8.

Figure 7-8. Locating the instrumentation key

152
Chapter 7 Getting Functions Production-Ready

Navigate to the Azure Functions and select Configuration, and then

Application Settings. Click “+ New application setting” and add a new key
APPLICATIONINSIGHTS_CONNECTION_STRING (see Figure 7-9).

Figure 7-9. Adding the key

The Azure Functions app is integrated with Application Insights, and

you can create custom telemetry events and other metrics in your Azure
app function.

Note You might find another application setting named

APPLICATIONINSIGHTS_INSTRUMENTATIONKEY. This setting has
been deprecated and will be removed in the future.

153
Chapter 7 Getting Functions Production-Ready

Disabling Built-in Logging

Because you have enabled Application Insights for your Azure Functions
app, it is imperative to disable the built-in logging that uses Azure Storage.
The built-in logging is good for lightweight workloads, such as testing in
lower environments but is not intended for use in production. The reason
for discouraging the use of built-in logging is that if the workload is high,
then the logs might be incomplete because of Azure Storage’s throttling.
To disable the built-in logging, you need to delete the
AzureWebJobsDashboard setting from the app settings. Just make sure that
this key is not being used in any other applications.

Configuring Categories and Log Levels

Application Insights is like a plug-and-play service for Azure Functions.
But when used with the default configuration, it could result in high-
volume data, and you may hit your data cap for Application Insights.
To avoid that, you can customize the configuration and ingest only the
required logs. To do that, you must first understand the log categories in
Azure Functions.

• The function runtime creates logs with a category that

begins with Host.

• The “Function started,” “function executed,” and

“function completed” logs have the Host.Executor
category.

• The logs that you write in your function have the

Function category.

154
Chapter 7 Getting Functions Production-Ready

You can configure which log levels can go to Application Insights for
these categories in the host.json file.

{
      "logging": {
            "fileLoggingMode": "always",
            "logLevel": {
                  "default": "Information",
                  "Host.Results": "Error",
                  "Function": "Error",
                  "Host.Aggregator": "Trace"
            }
      }
}

In the preceding configuration, you are specifying the following.

• For the Host.Results and Function categories,

you send logs with a log level of Error or higher to
Application Insights.

• For the Host.Aggregator category, you send logs with

Trace or Verbose and higher levels.
• For all other logs, you send them with a log level of
Information or higher.

So, now you are done, and your function is ready to be monitored
properly in production. Let’s see it in action in Figure 7-10.

155
Chapter 7 Getting Functions Production-Ready

Figure 7-10. Function in action

Monitoring functions in production is a necessity. It lets you monitor

load, errors, and requests and debug issues in production.

Securing Azure Functions

To make your functions production-ready, you must secure them to reduce
unauthorized access. In today’s world, securing your functions should be
one of the most important tasks, as there are a lot of data breaches, and
any data breach reduces people’s trust in the company and its websites. So,
it is paramount for you to secure Azure Functions.
The good thing about Azure Functions is that it provides an easy-to-use
configuration to secure your functions. Let’s go back to the HTTP-triggered
function you created in this book. Let’s copy the function URL, as shown in
Figure 7-11.

156
Chapter 7 Getting Functions Production-Ready

Figure 7-11. Copying the function

Copy this URL to a browser and add &name={provideYourName},

replacing YourName with any name, as shown in Figure 7-12.

Figure 7-12. Replacing YourName

As you can see, anyone who has your URL can access the function.
Since this is a basic function that does not interact with your database, it’s
OK. But consider a function like the OData API function you created in
Chapter 4. Now, if your endpoint is not secured (i.e., it does not require any
authentication/authorization), you invite hackers to easily get your data.
To avoid this, you must ensure that only authenticated users can access
your function. Let’s enable authentication/authorization for your function
using Active Directory.

157
Chapter 7 Getting Functions Production-Ready

Navigate to the function app you want to secure, click Authentication,

and click the “Add identity provider” button, as shown in Figure 7-13.

Figure 7-13. Adding Authentication

Choose an identity provider from the drop-down menu. Microsoft is

used for this example, as shown in Figure 7-14.

158
Chapter 7 Getting Functions Production-Ready

Figure 7-14. Setting “Action to take when request is not

authenticated”

On the next page, you can create a New AD app. Choose what
accounts are allowed to log in and how unauthenticated requests should
be handled. This creates the AD app and enables authentication/
authorization for this function, as shown in Figure 7-15.

159
Chapter 7 Getting Functions Production-Ready

Figure 7-15. Creating the AD app ID

Notice how Azure makes adding authentication to your function app

easy without writing any code to integrate with the identity provider.
Let’s use the URL in Figure 7-12. You see that now it asks you to log in
before showing the result, as shown in Figure 7-16.

160
Chapter 7 Getting Functions Production-Ready

Figure 7-16. Requesting a login

You have now secured your function, and only the users in your AD
tenant can access this function.

Configuring CORS on Azure Functions

In most cases where you want to use a function as an API, you are
running Azure Functions and your user interface or service that call Azure
Functions in different domains.
If that is the case, you must enable cross-origin site scripting (CORS) for
your function so that you can access it from different domains.
To do that, let’s follow these steps.
Let’s go back to the function and select CORS within the API section, as
shown in Figure 7-17.

161
Chapter 7 Getting Functions Production-Ready

Figure 7-17. Selecting CORS

Select Enable Access-Control-Allow-Credentials, as shown in

Figure 7-18. Let’s say you have an application running locally on http://
localhost:5000, and you want to access this function from this application.
In Allowed Origins, set this URL and click Save, as shown in Figure 7-18.

162
Chapter 7 Getting Functions Production-Ready

Figure 7-18. Setting the URL

You have enabled CORS on Azure Functions, and your function can
now be accessed from the domains you configured in Allowed Origins. To
enable all the URLs, set the Allowed Origins to *.

Summary
In this chapter, you learned about making your Azure Functions
production-ready by using the built-in logging capabilities, monitoring
using Application Insights, and securing your functions by configuring
authentication and CORS. The next chapter discusses some best practices
for working with Azure Functions.

163
CHAPTER 8

Best Practices for

Working with Azure
Functions
In this chapter, you learn about general best practices for working with
Azure Functions, including organizing your functions based on various
criteria, developing robust functions, and managing your Azure spending.

Organize Your Functions

Depending on the size and scale of your project, you could build several
functions to perform various operations. You can group them into one
or more function apps or deploy each function as a separate function
app. Before organizing your functions, you must consider aspects like
configuration, deployment, performance, scaling, and security of the
overall solution. Here are some commonly used patterns and practices.

Based on Business Function

Grouping functions based on a department or business function, such
as human resources and payroll, is a common practice among many
organizations. This pattern allows deployments and monitoring to be

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7_8
Chapter 8 Best Practices for Working with Azure Functions

independently carried out by the respective teams within the organization.

Each function app can be assigned to a Consumption plan or a Premium
plan and scaled individually according to the requirement.

For Performance and Scaling

A function should ideally be designed for short executions, performing
a small unit of work. This ensures that the function has a low memory
footprint. Several functions in the same app can still spike up the overall
memory usage. Dependencies for every function must be initialized and
loaded, which could lead to a slower app startup. This problem is also
referred to as the Azure Functions cold start problem and is apparent with
functions in the Consumption plan.
If you run multiple function apps in a single Premium plan or App
Service plan, these apps share the same resources allocated to the plan. If
you have one function app with a much higher memory requirement than
the other, it uses a large amount of memory on each instance to which the
app is deployed. Because this could leave less memory available for the
other apps on each instance, you might want to run a high-memory-using
function app like this in its own separate hosting plan.

By Configuration and Deployment

Functions within a function app inherit configuration from the host.json
file, which is used to configure the advanced behavior of function triggers
and the Azure Functions runtime. If you have functions that need similar
custom configurations, consider moving them into their own function app.
Functions that consume data from other functions or are dependent
on other functions to get triggered are good candidates to be grouped
together so that changes made to one function do not break functionality
on a dependent function and are always deployed together.

166
Chapter 8 Best Practices for Working with Azure Functions

By Privilege
Connection strings and other credentials stored in application settings
give all functions in the function app the same set of permissions in the
associated resource. Consider minimizing the number of functions with
access to specific credentials by moving functions that don’t use those
credentials to a separate function app. You can always use techniques
such as function chaining to pass data between functions in different
function apps.

By App Visibility

Azure Functions, especially HTTP-triggered public-facing APIs, have
different authentication, security, and availability requirements than
internal and private-facing APIs. A firewall usually protects public APIs. It
is a good practice to group such functions together.

Separate Function Apps for Each Function

While this pattern provides the most isolation from any other processes
or inter-service dependencies, it comes with a lot of operational overhead
because each function requires its own configuration, deployment,
and monitoring. Over time, the number of function apps could become
unmaintainable.
The Consumption plan pricing includes a monthly free grant of 1
million requests and 4,00,000 GB/s of resource consumption per month
per subscription in pay-as-you-go pricing across all function apps in that
subscription. You can deploy several function apps and still take advantage
of the pricing benefits provided by Azure.

167
Chapter 8 Best Practices for Working with Azure Functions

Develop Robust Functions

These are some important practices to consider while developing for scale
and performance.

Design for Statelessness

Functions should be designed to perform a small unit of work without
dependency on the outcome of a previous execution. For instance, a
function to email order summaries to customers should be able to resume
from where it was left in the previous occurrence. The function should rely
on the state of the order to determine if an email needs to be sent. Here,
the state belongs to the data being processed, not the function itself.
Leveraging a durable function when maintaining a state is necessary,
and there is a need to execute one function after the other in a particular
order across multiple functions.

Use Messaging Services

Use messaging services like Azure Queue Storage or Azure Service Bus
to relay messages between functions instead of invoking a function from
another function to transfer data. For instance, a function that processes
orders in an e-commerce system should place a message on a storage
queue with the data required by another function for sending a push
notification to the customer. Such a design allows the entire system to
scale out and scale when there is a sudden surge or decline in the number
of orders being processed. Messaging services also provide features like
delivery and ordering guarantees to ensure that the consuming function
receives and processes the message.

168
Chapter 8 Best Practices for Working with Azure Functions

Integrate with Monitoring Tools

The Azure portal provides basic metrics like memory utilization and
function execution count to monitor your functions. But these are
insufficient to determine if your app is experiencing performance issues
or throwing exceptions. APM (application performance monitoring) tools
like Application Insights can integrate easily with Azure Functions and
collect detailed information on function invocations, execution times, and
resource utilization and capture the stack trace of exceptions in your code
for troubleshooting. It also allows you to configure email and text alerts
when performance deteriorates beyond a certain threshold.
If you are already using other APM tools like AppDynamics and New
Relic, Azure Functions integrates with them to provide a consolidated
monitoring experience for all your applications.

Use Separate Storage Accounts

When you create an Azure Functions app, it also creates Azure storage
to store logs and execution schedules. This is metadata required by the
function to work. Data is being written to this storage account throughout
the function execution.
If you need to use storage blobs or queues with your function, you
should create separate storage accounts instead of reusing the one bound
to the function app. This is because storage accounts have their own
concurrency, throttling, and throughput limits and could seriously limit
the performance of your function app.

Account for Delays with Function Cold Starts

This problem is relevant to the Consumption plan. Due to the serverless
nature of Azure Functions, when functions are not invoked for an extended
duration (approx. 20 minutes), Azure deallocates the resources assigned to

169
Chapter 8 Best Practices for Working with Azure Functions

the app. Next time the function is triggered, the function’s runtime needs
to start and load dependencies before responding to the event. The total
time elapsed from the event occurring until the function responds to the
event is called cold start time (usually up to 10 seconds).
Implement a retry pattern in the calling app to wait and retry for a
response if it receives a time-out on the first attempt.
You can also create a function to periodically poll and invoke the other
functions to keep them “warm” even if they are not doing any actual work.
When a function needs to run, all the code is eventually loaded into
memory, which takes longer with a larger application. So, if the function
has many dependencies, the cold start times are longer due to increased
time for I/O operations from Azure Files and the longer time needed to
load them into memory.
For functions that cannot afford this latency, choose a Premium or App
Service plan so that the function’s host is always running.

Use Asynchronous Patterns

Make your code as asynchronous as possible. Avoid blocking calls and
ensure that any tasks, callbacks, threads, and processes initiated by the
function are completed before your function code returns. Functions don’t
track these background threads; site shutdown can occur regardless of
background thread status, which can cause unintended behavior in your
functions.
For example, suppose a function starts a background task and returns
a successful response before the task completes. In that case, the Azure
Functions runtime considers the execution as having been completed
successfully, regardless of the result of the background task. If this
background task is performing essential work, it may be preempted by site
shutdown, leaving that work in an unknown state.

170
Chapter 8 Best Practices for Working with Azure Functions

Manage Costs
Azure Functions’ instances in the Consumption plan automatically scale to
zero when there are no triggers for the functions. However, you are billed
for Application Insights depending on the volume of events collected. The
monitoring costs can become substantial and exceed the cost of Azure
Functions. Configure the log level to exclude Informational events if they
are not necessary for production. Application Insights also allows you to
specify sampling settings to reduce traffic and storage costs.
While inbound or ingress traffic to Azure is free, outbound or egress
traffic from Azure is billed. So, networking costs will apply if your Azure
Functions serve data to the outside world. Depending on the traffic
and volume of data transferred, these costs can add to your overall
application costs.
The storage account associated with the function is also charged for
the storage used. Although these charges are insignificant, they get added
to your total Azure spend.
Make sure that you understand the pricing model and closely monitor
your usage and Azure bill.

Summary
With this, you have come to the end of the book. We hope this book helped
you start your journey with building serverless applications on Azure. You
will learn more as you dig into the concepts explained in this book and
start building your own functions. Happy learning!

171
Index
A APPLICATIONINSIGHTS_
CONNECTION_STRING
Activity function, 81, 89, 92, 96, 98
key, 147
Activity trigger, 89
Application performance
message visibility, 90
monitoring (APM), 169
queues, 89
App Service plan, 9, 10,
return values, 90
135–143, 166
threading, 90
App Service (Slot), 121, 122
Admin authorization level, 21
Async HTTP APIs pattern, 84–86
Anonymous, 21, 64, 91, 105
Authenticated users, 157
Application insights, 171
Authorization-level function, 20
Azure Monitor, 147
Autoscaling, 97, 98
built-in logging, 154
Azure App Service, 6, 123
categories and log levels, Azure build providers, 124
configuration, 154–156 Azure data storage services, 110
CORS, 161–163 Azure DevOps account
integration, 147–150 bill, setup, 118
logging details, 147 link, subscription, 119
manual connection, 150–153 Azure DevOps configuration, 125
securing azure functions, Azure functions, 154, 170
147, 156–161 advanced tools, 25
Application insights, Azure application-level extensions, 26
functions app service plan, 9, 10
connection, 147 vs. Azure Logic Apps, 7, 8
manual connection, 152 vs. Azure WebJobs, 5, 6
APPLICATIONINSIGHTS_ cold start problem, 166
CONNECTION_ consumption plan, 9
STRING, 153 core tools, 34

R. Sawhney and K. Chanumolu, Beginning Azure Functions,
https://doi.org/10.1007/978-1-4842-9203-7
INDEX

Azure functions (cont.) B

features, 3, 4
Binding expressions
file hierarchy, 24
bindings, 32
logging-level extensions, 27
function app settings, 33
logging-level settings, 26
triggers, 32
microservices, 55
types, 33
output, 23
Bindings, 30, 31
Premium plan expressions patterns, 32, 33
benefits, 10 Blob storage–triggered function
billing, 10 C#
cold start problem, 11, 12 add function to
cold start vs. warm start, 11 workspace, 39
runtime host versions, 5 Azure logo, 36
runtime versions, 4 create new function, 36
serverless computing, 2, 3 .cs file, 40
Azure Logic Apps, 7, 8 function-v2-book, 39
Azure portal, 122 language, selection, 36
account creation, 14 local app setting, 38
create HTTP triggered namespace, 38
function, 20 naming function, 37
creation, 14–18 .NET runtime, selection, 37
function creation, 19, 22, 23 select storage account, Azure
name and settings, 16 functions runtime, 39
status checking, 18 storage, selection, 38
summary view, 17 subscription, selection, 38
Azure Queue Storage, 168 .template, selection, 37
Azure resource manager (ARM) host.json, 49
templates, 127, Node.js
128, 141–143 Azure storage account, 45
app service plan blob naming, 45
Azure functions, 136 code, 46, 49
Azure Service Bus, 4, 30, 168 function files, 46
Azure storage tables, 93, 94, 99 language, selection, 43
Azure WebJobs, 4–6, 69 naming function, 44

174
INDEX

template, selection, 44 Deployment Center, 122, 123, 142

workspace, addition, 45 Deployment functions
BlobTrigger, 30 App Service plan, 135–143
Browser-based interface, 3 ARM, 127, 128
Built-in logging, 145, 146 Azure DevOps account
setup, 115–120
code repository setup, 114, 115
C consumption plan, 128–131, 135
CI/CD pipeline, 141 continuous deployment
City_Travel function, 92 setup, 122–126
Client function, 80, 105 enabling deployment
Cold start time, 170 slots, 120–122
Command and query Deployments, 1
responsibility segregation Deployment slots, 121
(CQRS), 95 Durable functions
Concurrency throttling, 98 activity function, 80
Connection strings, 167 activity trigger, 89
Consumption plan, 9, 11, 15, application patterns
128–131, 135, 166, 167, 171 Async HTTP APIs
Continuous deployment, 4, 113 pattern, 84–86
Continuous integration/ fan-out/fan-in
continuous deployment pattern, 83, 84
(CI/CD), 113, 114 function chaining, 81, 82
Control queue, 93, 95 human intervention, 87–89
Cross-origin site scripting Monitoring, 86, 87
(CORS), 161–163 autoscaling, 97, 98
Custom configurations, 166 client function, 80
CustomerModel.cs file, 66 concurrency throttling, 98
creation, Azure Portal
activity, 107
D app details, 102
Dependent function, 166 App Service Editor,
Deploying functions 108, 109
CI/CD, 113, 114 create resource, 100

175
INDEX

Durable functions (cont.) Azure Functions Core

durable-func-new-book Tools, 34
function, 103 Visual Studio Tools, 34
Durable Functions HTTP
starter template, 104
”In-portal” environment, 103 F
installation, 105 Fan-out/fan-in pattern, 83, 84
Orchestrator_City function, File hierarchy, 24
106, 107 Kudu, 25
orchestrator client function, Function apps, 165, 166
105, 106 Azure portal
selecting orchestrator, 106 account creation, 14
templates, 104 create HTTP triggered
description, 79 function, 20
disaster recovery and geo creation, 14–18
distribution, 110, 111 function creation, 19, 22, 23
JSON object, 26, 89 item, 14, 16
orchestration client, 93, 94 name and settings, 16
orchestrator function, 80 status checking, 18
orchestrator function summary view, 17
replay, 98, 99 Function as a service (FaaS), 2
performance and scaling, 94–96 Function chaining, 81, 82
performance targets, 99, 100
stateful orchestration, 80
trigger bindings, 89
G
use case, 79 Geodistribution, 79, 110
Georeplication, 110
getSqlResults method, 72
E GetStatusAsync (.NET) API, 95
e-commerce system, 168 getStatus (JavaScript) API, 95
EventHubTrigger, 30 getStatus function, 85
Event sourcing pattern, 95 Gigabyte-seconds (GB-s), 9
Extensions, 32 GitHub Webhook, 30
installation Grouping functions, 165

176
INDEX

H I
History table, 94 ILogger, 105, 146
host.json file, 114, 155 Instance table, 95
HTTP Function boilerplate Integrated development
code, 21 environment (IDE), 3
HTTPTrigger, 30 Internal queue triggers, 95, 96
HttpTriggerCSharp.cs file, 68 Inter-service dependencies, 167
HTTP triggered function, Inventory management service, 55
20, 156 Isolated functions, 56
C#
create function, 62 J, K, L
CustomerModel.cs file, 66
JavaScript, 43, 146
folders creation, 66–68
language selection, 62
local URL, 70, 71 M
namespace, 64 Messaging services, 168
naming function, 62, 63 Microservices architecture, 54
NET 6.0 Isolated LTS benefits, 52, 53
runtime version, 63 nano services, 52
output, 70 trade-offs, 53
SqlClient package, 64, 65 Microsoft.Extensions.Logging, 146
template selection, 63 Monitoring functions, 145, 156
OData API, 71–77 Monitoring pattern, 86, 87
query output, OData API, 77 Monolithic applications
SQL server creation azure function, 54–56
connect to database, 60 Monolithic architecture
database creation, 58–60 APIs, 51
SQL database, vs. microservice
selection, 57 architecture, 52–54
tables, 61
HTTP-triggered public-facing N
APIs, 167
Nano services, 52
Human interaction, 87–89
Networking costs, 171

177
INDEX

O Q
OData API function, 157 QueueTrigger, 30
Open data protocol, 71
Orchestration trigger
activity function, 92 R
basic code, 92 Register binding extensions, 32
control queues, 93 Robust functions development
instances, 90 asynchronous patterns, 170
JSON object, 91 delays, function cold starts,
logic execution, 92 169, 170
message visibility, 91 messaging services, 168
queues, 90, 93 monitoring tools
return values, 91 integration, 169
single threading, 91 statelessness design, 168
OrchestrationTrigger, 92 storage accounts, 169
Orchestrator_City, 92 Routing strategy, 110
Orchestrator client, 93
Orchestrator function, 79, 80, 98, 99
Orchestrator scale-out, 96, 97
S
OrderFood function, 85 ScheduleConfiguration, 33
OrderProcess, 86 Seamless integration, 4
OrderProcess function, 85 Securing Azure functions
AD app ID, creation, 160
Serverless applications, 171
P Serverless computing, 2, 3
Performance targets, 99 Service bus trigger, 30
Platform features, 142 Single Premium plan, 166
Polyglot programming, 53 Software industry, 1
Premium plan, 10 SqlClient package, 65
Pricing model, 171 SqlClientHelper.cs file, 67
Programming languages, 4 SQLClientHelper.GetData
Publish, 15 method, 68

178
INDEX

SQL query, 76 GitHub Webhook, 30

SQL server, 57–61 HTTP, 30
SQL server management Queue, 30
studio, 61 Service Bus, 30
Storage account, 171 Timer, 30
Subscription, 15

V
T, U Virtual machine (VM), 9
TaskHubNameHistory, 94 Visual Studio, 34
Timers and compensation logic, 87 Visual Studio Code, 34
TimerTrigger, 30 Visual Studio Team
Traffic Manager, 110 Service (VSTS), 126
Trigger
Blob, 30
definition, 29 W, X, Y, Z
EventHub, 30 Work-item queue, 96

179

Microservices with Laravel - Martin Joo
100% (2)
Microservices with Laravel - Martin Joo
151 pages
Mastering Azure API Management - A Practical Approach To Designing and Implementing An API-Centric Enterprise Architecture
100% (2)
Mastering Azure API Management - A Practical Approach To Designing and Implementing An API-Centric Enterprise Architecture
273 pages
Design Patterns by Tutorials (2nd Edition) - 2018
100% (1)
Design Patterns by Tutorials (2nd Edition) - 2018
364 pages
Pravin Mishra - Cloud Computing With AWS - Everything You Need To Know To Be An AWS Cloud Practitioner-Apress
100% (2)
Pravin Mishra - Cloud Computing With AWS - Everything You Need To Know To Be An AWS Cloud Practitioner-Apress
340 pages
Anjani Kumar, Abhishek Mishra, Sanjeev Kumar - Architecting A Modern Data Warehouse For Large Enterprises - Build Multi-Cloud Modern Distributed Data Warehouses With Azure and AWS-Apress (2024)
No ratings yet
Anjani Kumar, Abhishek Mishra, Sanjeev Kumar - Architecting A Modern Data Warehouse For Large Enterprises - Build Multi-Cloud Modern Distributed Data Warehouses With Azure and AWS-Apress (2024)
378 pages
Kubernetes Basic To Advance End To End
100% (4)
Kubernetes Basic To Advance End To End
295 pages
Azure DevOps Explained Get Started With Azure DevOps and Develop Your DevOps Practices
100% (3)
Azure DevOps Explained Get Started With Azure DevOps and Develop Your DevOps Practices
497 pages
Pro Microservices in .NET 6 by Sean Whitesell, Et Al (2022)
No ratings yet
Pro Microservices in .NET 6 by Sean Whitesell, Et Al (2022)
310 pages
1.2 Beginning Microsoft Dataverse - Exploiting Microsoft's Low-Code
100% (3)
1.2 Beginning Microsoft Dataverse - Exploiting Microsoft's Low-Code
239 pages
Getting Started With Angular (2023) Create and Deploy Angular App
No ratings yet
Getting Started With Angular (2023) Create and Deploy Angular App
410 pages
Beginning Ansible Concepts and Application Provisioning, Configuring, and Managing Servers, Applications, and Their... (Shaun Smith, Peter Membrey)
No ratings yet
Beginning Ansible Concepts and Application Provisioning, Configuring, and Managing Servers, Applications, and Their... (Shaun Smith, Peter Membrey)
296 pages
Infrastructure Leader's Guide To Google Cloud
50% (2)
Infrastructure Leader's Guide To Google Cloud
366 pages
Continuous Integration (CI) and Continuous Delivery (CD)
100% (1)
Continuous Integration (CI) and Continuous Delivery (CD)
429 pages
Azure DevOps Explained
75% (4)
Azure DevOps Explained
438 pages
Azure Devops Explained
88% (8)
Azure Devops Explained
438 pages
Hands-On Kubernetes On Azure
100% (4)
Hands-On Kubernetes On Azure
330 pages
Kubernetes User Roadmap
No ratings yet
Kubernetes User Roadmap
20 pages
Rob Reid - Practical CockroachDB - Building Fault-Tolerant Distributed SQL Databases-Apress (2022)
No ratings yet
Rob Reid - Practical CockroachDB - Building Fault-Tolerant Distributed SQL Databases-Apress (2022)
264 pages
Data Integration Life Cycle Management With SSIS - A Short Introduction by Example
100% (1)
Data Integration Life Cycle Management With SSIS - A Short Introduction by Example
188 pages
Kafka in Action
100% (3)
Kafka in Action
209 pages
Azure Devops Complete Ci CD Pipeline PDF
100% (8)
Azure Devops Complete Ci CD Pipeline PDF
109 pages
Java Book
No ratings yet
Java Book
97 pages
The Optimal RabbitMQ Guide
No ratings yet
The Optimal RabbitMQ Guide
138 pages
Securing Microservice APIs
No ratings yet
Securing Microservice APIs
51 pages
The Ultimate C - C - TS4FI - 1909 - SAP Certified Application Associate - SAP S4HANA For Financial Accounting Associates (SAP S4HANA 1909)
No ratings yet
The Ultimate C - C - TS4FI - 1909 - SAP Certified Application Associate - SAP S4HANA For Financial Accounting Associates (SAP S4HANA 1909)
2 pages
Unit-2 Domain Specific Iots
No ratings yet
Unit-2 Domain Specific Iots
50 pages
POST TEST St.-Louis-Reviewer
No ratings yet
POST TEST St.-Louis-Reviewer
239 pages
A Complete Guide To DevOps With AWS
100% (1)
A Complete Guide To DevOps With AWS
579 pages
Dokumen - Pub - Microsoft Azure in Action Meap v06
No ratings yet
Dokumen - Pub - Microsoft Azure in Action Meap v06
341 pages
Shimon Ifrah - Getting Started With Containers in Azure - Deploy Secure Cloud Applications Using Terraform-Apress (2024)
No ratings yet
Shimon Ifrah - Getting Started With Containers in Azure - Deploy Secure Cloud Applications Using Terraform-Apress (2024)
221 pages
Microsoft Azure For Java Developers D.
No ratings yet
Microsoft Azure For Java Developers D.
361 pages
AZ 900 2022 OCT Shrunk New
No ratings yet
AZ 900 2022 OCT Shrunk New
405 pages
Docker in Action - Manning (2016)
100% (3)
Docker in Action - Manning (2016)
306 pages
Architecting Cloud Native NET Apps For Azure
100% (1)
Architecting Cloud Native NET Apps For Azure
197 pages
Azure Cosmos DB
No ratings yet
Azure Cosmos DB
835 pages
Apache Kafka Transaction Data Streaming For Dummies
No ratings yet
Apache Kafka Transaction Data Streaming For Dummies
47 pages
Azure Virtual Desktop Specialist Exam Study Guide - AZ-140 (Shabaz Darr)
No ratings yet
Azure Virtual Desktop Specialist Exam Study Guide - AZ-140 (Shabaz Darr)
257 pages
Mastering Docker A Comprehensive Guide
No ratings yet
Mastering Docker A Comprehensive Guide
105 pages
Azure For Developers
100% (3)
Azure For Developers
65 pages
The Enterprise Linux Administrator
No ratings yet
The Enterprise Linux Administrator
607 pages
Containerized Docker Application Lifecycle With Microsoft Platform and Tools
100% (1)
Containerized Docker Application Lifecycle With Microsoft Platform and Tools
104 pages
Azure Book
67% (3)
Azure Book
130 pages
YAML-its Usage in Kubernetes
100% (1)
YAML-its Usage in Kubernetes
46 pages
Warner AZ400 Slides
No ratings yet
Warner AZ400 Slides
232 pages
GCP Pde
100% (3)
GCP Pde
200 pages
Reed, Mark - Kubernetes - The Ultimate Beginners Guide To Effectively Learn Kubernetes Step-by-Step-Publishing Factory (2020)
No ratings yet
Reed, Mark - Kubernetes - The Ultimate Beginners Guide To Effectively Learn Kubernetes Step-by-Step-Publishing Factory (2020)
88 pages
THE Definitive Cheat Sheet: Kubectl
No ratings yet
THE Definitive Cheat Sheet: Kubectl
23 pages
TheNewStack Book1 2 TheStateOfTheKubernetesEcosystem SecondEdition PDF
No ratings yet
TheNewStack Book1 2 TheStateOfTheKubernetesEcosystem SecondEdition PDF
81 pages
Advanced Office Add-In Development
No ratings yet
Advanced Office Add-In Development
40 pages
Kubernetes Administrator Roadmap
No ratings yet
Kubernetes Administrator Roadmap
22 pages
Kubernetes II Deep Dive
No ratings yet
Kubernetes II Deep Dive
74 pages
Introduction To Azure Cosmos DB PDF
No ratings yet
Introduction To Azure Cosmos DB PDF
1,816 pages
Quarkus Fascicle Practising v1
No ratings yet
Quarkus Fascicle Practising v1
210 pages
Learn Kubernetes 5 Minutes at A Time
No ratings yet
Learn Kubernetes 5 Minutes at A Time
187 pages
Ansible For Teenagers
No ratings yet
Ansible For Teenagers
23 pages
Terraform For Teenagers
No ratings yet
Terraform For Teenagers
27 pages
Angular Succinctly
No ratings yet
Angular Succinctly
124 pages
Apress Ansible For Kubernetes by Example
100% (2)
Apress Ansible For Kubernetes by Example
300 pages
Vue - Js Succinctly
No ratings yet
Vue - Js Succinctly
101 pages
Azure Architecturing PDF
No ratings yet
Azure Architecturing PDF
113 pages
Documentacao Langchain
No ratings yet
Documentacao Langchain
53 pages
Cloud Migration Mastery: Complete Guide To Seamless Cloud Integration With AWS, Microsoft Azure, VMware & NaviSite
From Everand
Cloud Migration Mastery: Complete Guide To Seamless Cloud Integration With AWS, Microsoft Azure, VMware & NaviSite
Rob Botwright
No ratings yet
Download ebooks file Beginning Azure Functions: Building Scalable and Serverless Apps 2nd Edition Rahul Sawhney all chapters
100% (1)
Download ebooks file Beginning Azure Functions: Building Scalable and Serverless Apps 2nd Edition Rahul Sawhney all chapters
41 pages
Beginning Azure Functions: Building Scalable and Serverless Apps 2nd Edition Rahul Sawhney - The complete ebook version is now available for download
100% (1)
Beginning Azure Functions: Building Scalable and Serverless Apps 2nd Edition Rahul Sawhney - The complete ebook version is now available for download
55 pages
Ds Dragon Professional v16 en Us
No ratings yet
Ds Dragon Professional v16 en Us
2 pages
极点分配
100% (2)
极点分配
8 pages
CPP Presentation
No ratings yet
CPP Presentation
15 pages
Report
No ratings yet
Report
9 pages
Dirichlet series
No ratings yet
Dirichlet series
13 pages
Cybersecurity Hacker Linux With Little Trouble Cybercrime and The
No ratings yet
Cybersecurity Hacker Linux With Little Trouble Cybercrime and The
580 pages
NVIDIA GPU CUDA Toolkit & Driver Installation in Linux OS
100% (1)
NVIDIA GPU CUDA Toolkit & Driver Installation in Linux OS
2 pages
Schneider Electric - EcoStruxure-Security-Expert - SX-SRVR
No ratings yet
Schneider Electric - EcoStruxure-Security-Expert - SX-SRVR
2 pages
TrainingDocument-E01001 D4 ENDEAVOR
No ratings yet
TrainingDocument-E01001 D4 ENDEAVOR
2 pages
02mindset1 VocabGram MorePrac2
100% (1)
02mindset1 VocabGram MorePrac2
4 pages
Greesh Project Report
No ratings yet
Greesh Project Report
59 pages
Dannyokec Adsense Course Terms and Conditions
No ratings yet
Dannyokec Adsense Course Terms and Conditions
12 pages
Milestone Innovative Academy: Monthly Examination in Computer 2
No ratings yet
Milestone Innovative Academy: Monthly Examination in Computer 2
2 pages
Important 16 Mark Questions-19!1!2024
No ratings yet
Important 16 Mark Questions-19!1!2024
5 pages
Hfe Yamaha Dsp-A595a Service PDF
No ratings yet
Hfe Yamaha Dsp-A595a Service PDF
58 pages
Current Log
No ratings yet
Current Log
23 pages
Exhibit 04 System Completion
No ratings yet
Exhibit 04 System Completion
9 pages
FortiAnalyzer 7.0.1 Administration Guide
No ratings yet
FortiAnalyzer 7.0.1 Administration Guide
353 pages
Writeup Infinity
No ratings yet
Writeup Infinity
5 pages
CV 22
No ratings yet
CV 22
1 page
Monetization SBLC
100% (3)
Monetization SBLC
7 pages
API Docs
No ratings yet
API Docs
13 pages
Barbara Pareglio Amaia White GSMA MIoT
No ratings yet
Barbara Pareglio Amaia White GSMA MIoT
38 pages
Mawjuud - SEO Site Checkup - 18-07-2024
No ratings yet
Mawjuud - SEO Site Checkup - 18-07-2024
23 pages
Eaton Y7 197212 EASY E4 UC 12RCX1 en - GB
No ratings yet
Eaton Y7 197212 EASY E4 UC 12RCX1 en - GB
9 pages
Picture Book Homework
100% (1)
Picture Book Homework
6 pages