How can states be preserved between stages and jobs within a single stage?

Artifacts are typically used to preserve state between stages in CI/CD pipelines.

Artifacts are files or directories created by one stage or job that can be used by other stages or jobs later.

Artifacts can also be used to maintain state between jobs within a single stage. Some CI/CD systems allow passing the output of one job as input to another job within the same stage.

Implementation examples:

• Jenkins: Uses the artifact mechanism (archiveArtifacts) to save data between stages. To pass data between jobs within a stage, plugins or scripts can be used to save data to temporary files, which are then read by the next job.
• GitLab CI: Uses the artifacts keyword to define artifacts accessible in subsequent stages. To pass data between jobs within a single stage, artifacts with limited lifespan or intermediate files can also be used.
• GitHub Actions: Uses upload-artifact and download-artifact to save and retrieve artifacts between jobs and workflows (which are similar to stages).

Example of using artifacts in GitLab CI:

stages:
  - build
  - test

build_job:
  stage: build
  script:
    - echo "Building the application..."
    - mkdir build_output
    - echo "Build data" > build_output/data.txt
  artifacts:
    paths:
      - build_output/

test_job:
  stage: test
  script:
    - echo "Testing the application..."
    - cat build_output/data.txt # Accessing artifact from build_job
  dependencies:
    - build_job