DevOps / CI/CD pipelines

CI/CD

• Continuous Integration (CI)
  • fast integration of changes, avoiding long-living branches
• Continuous Delivery (CD)
  • automation of release process, including testing and deployment
  • deployment pipeline
• Continuous Deployment
  • actually deploying every change in production automatically

CI/CD Services

• Tools for automate deployment pipeline
• Build, test, deploy
• Examples: GitHub Actions, GitLab CI, Jenkins, TeamCity

GitLab CI/CD

GitLab CI/CD

• CI/CD pipeline automation integrated in GitLab
  • menu: Build ⇨ Pipelines
• Run pipelines on every push, merge request, or schedule
• Configuration:
  • .gitlab-ci.yml file in the root of the repository

YAML

• Data serialization language
• Writable and readable by humans
• Structure is defined by indentation

key: value
nested:
  person:
    name: John
    age: 30
list:
  - item 1
  - item 2
inline-list: [item 1, item 2]

strings: "hello world"
quotation-marks: not required
numbers: 42
booleans:
  - true
  - false
  - yes
  - no
# Comments

Caveats

• Indentation is important
• No tabs allowed, only spaces
• Quotation for strings is sometimes required:
  • special characters (e.g. :, #, -)
  • numbers
  • booleans (e.g. yes, no, true, false)

.gitlab-ci.yml

Stages

• Define the order of execution of jobs

stages:
  - build
  - test
  - deploy

Jobs

• Specify tasks to be executed

compile:
  stage: build
  script:
    - echo "Compiling the code..."
    - gcc main.c -o main

How the job is executed?

GitLab Runners

• Separate services executing the jobs
• Shared or specific for a project
• Steps:
  1. Clone the repository
  2. Execute the script
  3. Report the results
• Can use Docker images for execution environment

Docker Image

compile:
  stage: build
  image: gcc:latest
  script:
    - echo "Compiling the code..."
    - gcc main.c -o main

What with the results?

Status of the job

• Success (green)
• Failed (red)
• Based on the exit code of the script
• Example failures: compilation errors, failed tests, etc.
• Next stages are not executed if a job fails

Artifacts

• Useful results of the job that we want to keep
• Files or directories
• Available in GitLab UI
• Can be used in jobs in later stages

Artifacts

compile:
  stage: build
  image: gcc:latest
  script: 
    - gcc main.c -o main
  artifacts:
    paths:
      - main

Default settings

default:
  image: python:3.9  # default image for all jobs
  before_script:     # run before each job
    - pip install -r requirements.txt

CI/CD variables

• Predefined by GitLab
  • e.g. CI_COMMIT_REF_NAME, CI_PIPELINE_ID, etc.
• Defined in the .gitlab-ci.yml (variables section)
• Defined in GitLab project settings
  • Settings ⇨ CI/CD ⇨ Variables

How to use variables?

variables:
  PRODUCTION_SERVER: "prod.example.com"
deploy:
  stage: deploy
  script:
    - echo "Deploying to $PRODUCTION_SERVER server..."
    - scp main user@$PRODUCTION_SERVER:/path/to/deploy

Secrets

• Sensitive information (e.g. API keys, passwords, etc.)
• Should not be stored in the repository
• Use GitLab CI/CD variables with “Protected” and “Masked” options enabled
• Accessed as regular environment variables

Conclusion

• CI/CD pipelines automate the software delivery process
• GitLab CI/CD is one of the popular tools
  • we only scratched the surface
• Other tools use similar concepts, just different syntax