Azure Devops Solution

[reboottime]

Overview

Revisit Azure DevOps as a software development automation solution, to have a systematic understanding.

Other than providing a higher overview of Azure Devops platform, this article also delves the main concepts of Azure pipeline, including

• the yaml file basics for azure pipeline
• and practices about how to connect services to azure devops
• the strategy on code deployment.

This article is built upon this tutorial

FAQs:

• What is Devops: Devops is a kind of solution to automate (fully or mostly of) software development cycle.

• Why Azure Devops for small agile team: Azure Devops solution contains the whole life cycle management of software development,
  • project management
  • code pipeline management
  • Code base management, azure repository
  • and code deployment

[reboottime]

Azure Pipeline

Azure supports yaml language to define your code pipeline, which means your pipeline can be part of your code.

There are some main concepts in Azure pipeline

• jobs
• steps, can either be represented as script or task
• pool
• agent
• artifacts

We will explore those concepts by use applications.

The Basics Example

The basic syntax of azure pipeline configuration

• trigger: at what condition triggers a pipeline run, for example

trigger:
- main

• steps: step is the main building block of azure pipeline, and

  • what does it do: what steps are there for your pipeline, such as running test, pack packages and publish your packed images, etc.
  • the presentation of steps
    • can either be a task: a pre-created script created to offer convenience
    • or be a script: you write your script for each step

Bellow is an example of pipeline configuration using yaml language

[reboottime]

Azure Pipeline Jobs

1. When you will want to use jobs in pipeline configuration

In some scenarios, you want to test your application on different environments, for example, test your application in Linux and Windows environment.

Yet you do not want to wait for too long for the pipeline to run through, the solution is to define jobs such that your pipeline can be run on two different environment in parallel.

For the above scenario, we will introduce three main concepts

• Jobs: a standalone execution boundary for a specific environment that you want your pipeline steps to run through.
• Pool: What kind of services( infrastructure) your agent runs on, for example, the windows system infrastructure, and the linux infrastructure
• Agent: The service that runs the job you defined in your pipeline.

2. Illustrations on using job

VS

For the above example( the second screenshot), each job runs on a different environment and has its own step(s)

[reboottime]

Azure Artifacts

Artifacts are the deliverable file(s) that we use to publish our (part of) application(s).

Artifacts differ based on the programming language used for writing the application, for example,

• java language: jar file
• Microsoft .net: nudge
• Javascript: Zipped javascript file(s)

we use zipped file(s) and upload the application deployment environment and unzip the file then deploy the app. This step involves in archive the built artifacts using @ArchiveFiles@1 task, for example, archive the a website's landing page output artifacts built with nextjs SSR code.

- task: ArchiveFiles@1
  displayName: Generate landing page build package
  inputs:
    rootFolder: '$(Build.SourcesDirectory)/landing/out/'
    includeRootFolder: false
    archiveType: 'default'
    archiveFile: '$(Build.ArtifactStagingDirectory)/landing.zip'
    replaceExistingArchive: true

However, in the modern software development we use Docker as artifacts

The container registry way

[reboottime]

Stages in Azure pipeline

• The CI/CD process illustration

• The yaml code that represents the above process

---

What is stage

• State is a logical boundary in the pipeline
• Each stage contains 1 or more jobs
• By default, they run one after the other

[reboottime]

Each stage are different, for example, in deployment stage, we do not check out source code from git.

Strategies to deploy code

Usually, we do not deploy code directly to production

We test the code gradually and deploy code gradually.

Based on the above strategy, we can have different stages in our pipeline, for example, as follows

[reboottime]

Templates in Azure Pipelines

Template is a reusable configuration that you can reuse it in different stages, for example

[reboottime]

Azure supports create environment for you for different purposes, here is where you can find the UI exposure to create environment manually

Then you can configure your azure pipeline based on different enviroments

[reboottime]

Services connections

The cases that we want to connect services to azure Devops workflow

Azure has UI exposures to support add credential connections manually