Introduction
In the dynamic world of software development, the right implementation of DevOps practices can make or break a project. Microsoft Azure DevOps offers an impressive set of tools to efficiently build Continuous Integration and Continuous Deployment (CI/CD) pipelines. In this article, we take a look at a few helpful fundamentals.
YAML Templates and Repository Resources
YAML Templates provide a reusable structure for Azure DevOps Pipelines, allowing pipeline configurations to be managed as code. In combination with Repository Resources, these templates can be defined once and then used across multiple projects. The templates live in one Git repository, while the individual projects reside in separate repositories. This ensures that all projects use consistent and up-to-date pipeline configurations.
resources:
repositories:
- repository: templates
type: git
endpoint: # Name of the service connection in Azure DevOps
name: # Name of the repository containing the templates
ref: # Optional, git ref in the target repository
stages:
- template: templates/deploy-template.yml@templates
Automating Deployments with Pipeline Resources
Azure DevOps enables deployment automation through Build Completion Triggers, which ensure deployments are only performed when the previous build was successful. Pipeline Resources and their trigger configuration allow fine-grained control over when the current pipeline should be started automatically.
For example, a deployment can be triggered automatically as soon as the build of an application component has succeeded.
resources:
pipelines:
- pipeline: BuildPipeline
source: BuildPipeline
trigger:
branches:
include:
- main
stages:
# Deployment Stages
Multi-Stage Pipelines
Pipelines represent multi-stage processes, where stages can represent — for example — the build of a specific application component or a deployment to a defined system. The dependsOn attribute allows complex dependencies between different stages of a pipeline to be expressed, enabling precise control over the deployment flow across different environments.
stages:
- stage: deploy_to_staging
jobs:
- job: ...
- stage: deploy_to_environment_a
# Only runs when "staging" has been deployed successfully
dependsOn: deploy_to_staging
condition: succeeded()
jobs:
- job: ...
- stage: deploy_to_environment_a
dependsOn: deploy_to_staging
condition: succeeded()
jobs:
- job: ...
Summary and Outlook
Azure DevOps offers a wide range of features to design modern software development processes efficiently and securely. The features shown here form the foundation for a sustainable design of automated CI/CD pipelines.
But there is much more that Azure DevOps Pipelines can do. In the next articles on this topic, we will look at how deployments can be secured through approvals, how to handle secrets ideally, and how pipelines can be used to automate deployments to Kubernetes with Helm.