CI/CD Pipelines and DevOps Principles

What it is

A CI/CD pipeline is an automated path that takes a code change from a commit to a running application.

CI stands for continuous integration, and CD stands for continuous delivery or continuous deployment. DevOps is the wider way of working that makes such pipelines natural, by joining development and operations.

The aim is to release small changes often, with fewer errors and less manual work.

How it works

A pipeline is a series of stages, and a failure in one stage stops the change from moving on.

Continuous integration

Developers merge their work into a shared branch often, for example several times a day. Each merge triggers an automatic build and a set of automated tests.

The benefit is quick feedback. A broken change is found within minutes, while the developer still remembers it.

Continuous delivery and continuous deployment

Pipeline components

A pipeline is usually described in a file kept in the repository, a practice called pipeline as code. This is a simple example in the style of GitLab CI.

stages:
  - build
  - test
  - deploy

build_job:
  stage: build
  script:
    - docker build -t myapp:1.0 .

test_job:
  stage: test
  script:
    - python -m unittest discover

deploy_job:
  stage: deploy
  script:
    - ./deploy.sh staging

The stages list sets the order. Each job names its stage and lists the commands in script. Jenkins, GitHub Actions and GitLab CI follow the same idea with different file formats.

Environments and release strategies

Changes often flow through several environments: development, test, staging and production. Each one is closer to the real thing.

Teams may reduce risk with a blue green release, which keeps two copies of production and switches users from one to the other, or a canary release, which sends a small share of users to the new version first.

Benefits

• Faster releases in small steps.
• Fewer manual mistakes.
• The same steps run the same way each time.
• Early detection of faults.
• An easy way to roll back to an earlier artifact.

DevOps principles

DevOps joins development and operations so that they share responsibility from idea to running service. Its main principles are:

• Collaboration. Teams share goals and tools, and remove the wall between building and running.
• Automation. Builds, tests, deployments and infrastructure are scripted and repeatable.
• Continuous integration and delivery. Small changes flow to production often.
• Monitoring and feedback. Metrics and logs show how the service behaves, and the findings guide the next change.
• Continuous improvement. Teams learn from failures without blame and refine their process.
• Infrastructure as code. Environments are described in files and kept under version control.

DevOps grew out of agile and lean ideas. The common summary is that you build it, and you run it.