复制的官方GitLab 文档

stage: Verify group: Continuous Integration info: To determine the technical writer assigned to the Stage/Group associated with this page, see https://about.gitlab.com/handbook/engineering/ux/technical-writing/#assignments type: reference

Keyword reference for the .gitlab-ci.yml file (FREE)

This document lists the configuration options for your GitLab .gitlab-ci.yml file.

• For a quick introduction to GitLab CI/CD, follow the quick start guide.
• For a collection of examples, see GitLab CI/CD Examples.
• To view a large .gitlab-ci.yml file used in an enterprise, see the .gitlab-ci.yml file for gitlab.

When you are editing your .gitlab-ci.yml file, you can validate it with the CI Lint tool.

Job keywords

A job is defined as a list of keywords that define the job’s behavior.

The keywords available for jobs are:

Unavailable names for jobs

You can’t use these keywords as job names:

• image
• services
• stages
• types
• before_script
• after_script
• variables
• cache
• include

Custom default keyword values

You can set global defaults for some keywords. Jobs that do not define one or more of the listed keywords use the value defined in the default: section.

These job keywords can be defined inside a default: section:

• after_script
• artifacts
• before_script
• cache
• image
• interruptible
• retry
• services
• tags
• timeout

The following example sets the ruby:2.5 image as the default for all jobs in the pipeline. The rspec 2.6 job does not use the default, because it overrides the default with a job-specific image: section:

default:
  image: ruby:2.5

rspec:
  script: bundle exec rspec

rspec 2.6:
  image: ruby:2.6
  script: bundle exec rspec

Global keywords

Some keywords are not defined in a job. These keywords control pipeline behavior or import additional pipeline configuration:

stages

Use stages to define stages that contain groups of jobs. stages is defined globally for the pipeline. Use stage in a job to define which stage the job is part of.

The order of the stages items defines the execution order for jobs:

• Jobs in the same stage run in parallel.
• Jobs in the next stage run after the jobs from the previous stage complete successfully.

For example:

stages:
  - build
  - test
  - deploy

1. All jobs in build execute in parallel.
2. If all jobs in build succeed, the test jobs execute in parallel.
3. If all jobs in test succeed, the deploy jobs execute in parallel.
4. If all jobs in deploy succeed, the pipeline is marked as passed.

If any job fails, the pipeline is marked as failed and jobs in later stages do not start. Jobs in the current stage are not stopped and continue to run.

If no stages are defined in the .gitlab-ci.yml file, then build, test and deploy are the default pipeline stages.

If a job does not specify a stage, the job is assigned the test stage.

To make a job start earlier and ignore the stage order, use the needs keyword.

workflow

Introduced in GitLab 12.5

Use workflow: to determine whether or not a pipeline is created. Define this keyword at the top level, with a single rules: keyword that is similar to rules: defined in jobs.

You can use the workflow:rules templates to import a preconfigured workflow: rules entry.

workflow: rules accepts these keywords:

• if: Check this rule to determine when to run a pipeline.
• when: Specify what to do when the if rule evaluates to true.
  • To run a pipeline, set to always.
  • To prevent pipelines from running, set to never.

When no rules evaluate to true, the pipeline does not run.

Some example if clauses for workflow: rules:

See the common if clauses for rules for more examples.

In the following example, pipelines run for all push events (changes to branches and new tags). Pipelines for push events with -draft in the commit message don’t run, because they are set to when: never. Pipelines for schedules or merge requests don’t run either, because no rules evaluate to true for them:

workflow:
  rules:
    - if: $CI_COMMIT_MESSAGE =~ /-draft$/
      when: never
    - if: '$CI_PIPELINE_SOURCE == "push"'

This example has strict rules, and pipelines do not run in any other case.

Alternatively, all of the rules can be when: never, with a final when: always rule. Pipelines that match the when: never rules do not run. All other pipeline types run:

workflow:
  rules:
    - if: '$CI_PIPELINE_SOURCE == "schedule"'
      when: never
    - if: '$CI_PIPELINE_SOURCE == "push"'
      when: never
    - when: always

This example prevents pipelines for schedules or push (branches and tags) pipelines. The final when: always rule runs all other pipeline types, including merge request pipelines.

If your rules match both branch pipelines and merge request pipelines, duplicate pipelines can occur.

workflow:rules templates

Introduced in GitLab 13.0.

GitLab provides templates that set up workflow: rules for common scenarios. These templates help prevent duplicate pipelines.

The Branch-Pipelines template makes your pipelines run for branches and tags.

Branch pipeline status is displayed in merge requests that use the branch as a source. However, this pipeline type does not support any features offered by merge request pipelines, like pipelines for merge results or merge trains. This template intentionally avoids those features.

To include it:

include:
  - template: 'Workflows/Branch-Pipelines.gitlab-ci.yml'

The MergeRequest-Pipelines template makes your pipelines run for the default branch, tags, and all types of merge request pipelines. Use this template if you use any of the the pipelines for merge requests features.

To include it:

include:
  - template: 'Workflows/MergeRequest-Pipelines.gitlab-ci.yml'

Switch between branch pipelines and merge request pipelines

Introduced in GitLab 13.8.

To make the pipeline switch from branch pipelines to merge request pipelines after a merge request is created, add a workflow: rules section to your .gitlab-ci.yml file.

If you use both pipeline types at the same time, duplicate pipelines might run at the same time. To prevent duplicate pipelines, use the CI_OPEN_MERGE_REQUESTS variable.

The following example is for a project that runs branch and merge request pipelines only, but does not run pipelines for any other case. It runs:

• Branch pipelines when a merge request is not open for the branch.
• Merge request pipelines when a merge request is open for the branch.

workflow:
  rules:
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event"'
    - if: '$CI_COMMIT_BRANCH && $CI_OPEN_MERGE_REQUESTS'
      when: never
    - if: '$CI_COMMIT_BRANCH'

If the pipeline is triggered by:

• A merge request, run a merge request pipeline. For example, a merge request pipeline can be triggered by a push to a branch with an associated open merge request.
• A change to a branch, but a merge request is open for that branch, do not run a branch pipeline.
• A change to a branch, but without any open merge requests, run a branch pipeline.

You can also add a rule to an existing workflow section to switch from branch pipelines to merge request pipelines when a merge request is created.

Add this rule to the top of the workflow section, followed by the other rules that were already present:

workflow:
  rules:
    - if: $CI_COMMIT_BRANCH && $CI_OPEN_MERGE_REQUESTS && $CI_PIPELINE_SOURCE == "push"
      when: never
    - ...                # Previously defined workflow rules here

Triggered pipelinesthat run on a branch have a$CI_COMMIT_BRANCH

include

Moved to GitLab Free in 11.4.

Use include to include external YAML files in your CI/CD configuration. You can break down one long gitlab-ci.yml file into multiple files to increase readability, or reduce duplication of the same configuration in multiple places.

You can also store template files in a central repository and include them in projects.

include requires the external YAML file to have the extensions .yml or .yaml, otherwise the external file is not included.

You can’t use YAML anchors across different YAML files sourced by include. You can only refer to anchors in the same file. To reuse configuration from different YAML files, use !reference tags or the extends keyword.

include supports the following inclusion methods:

When the pipeline starts, the .gitlab-ci.yml file configuration included by all methods is evaluated. The configuration is a snapshot in time and persists in the database. GitLab does not reflect any changes to the referenced .gitlab-ci.yml file configuration until the next pipeline starts.

The include files are:

• Deep merged with those in the .gitlab-ci.yml file.
• Always evaluated first and merged with the content of the .gitlab-ci.yml file, regardless of the position of the include keyword.

NOTE: Use merging to customize and override included CI/CD configurations with local configurations. Local configurations in the .gitlab-ci.yml file override included configurations.

Variables with include (FREE SELF)

• Introduced in GitLab 13.8.
• Feature flag removed in GitLab 13.9.

You can use some predefined variables in include sections in your .gitlab-ci.yml file:

include:
  project: '$CI_PROJECT_PATH'
  file: '.compliance-gitlab-ci.yml'

For an example of how you can include these predefined variables, and the variables’ impact on CI/CD jobs, see this CI/CD variable demo.

include:local

Use include:local to include a file that is in the same repository as the .gitlab-ci.yml file. Use a full path relative to the root directory (/).

If you use include:local, make sure that both the .gitlab-ci.yml file and the local file are on the same branch.

You can’t include local files through Git submodules paths.

All nested includes are executed in the scope of the same project, so it’s possible to use local, project, remote, or template includes.

Example:

include:
  - local: '/templates/.gitlab-ci-template.yml'

You can also use shorter syntax to define the path:

include: '.gitlab-ci-production.yml'

Use local includes instead of symbolic links.

include:file

Introduced in GitLab 11.7.

To include files from another private project on the same GitLab instance, use include:file. You can use include:file in combination with include:project only. Use a full path, relative to the root directory (/).

For example:

include:
  - project: 'my-group/my-project'
    file: '/templates/.gitlab-ci-template.yml'

You can also specify a ref. If you do not specify a value, the ref defaults to the HEAD of the project:

include:
  - project: 'my-group/my-project'
    ref: master
    file: '/templates/.gitlab-ci-template.yml'

  - project: 'my-group/my-project'
    ref: v1.0.0
    file: '/templates/.gitlab-ci-template.yml'

  - project: 'my-group/my-project'
    ref: 787123b47f14b552955ca2786bc9542ae66fee5b  # Git SHA
    file: '/templates/.gitlab-ci-template.yml'

All nested includes are executed in the scope of the target project. You can use local (relative to target project), project, remote, or template includes.

Multiple files from a project

• Introduced in GitLab 13.6.
• Feature flag removed in GitLab 13.8.

You can include multiple files from the same project:

include:
  - project: 'my-group/my-project'
    ref: master
    file:
      - '/templates/.builds.yml'
      - '/templates/.tests.yml'

include:remote

Use include:remote with a full URL to include a file from a different location. The remote file must be publicly accessible by an HTTP/HTTPS GET request, because authentication in the remote URL is not supported. For example:

include:
  - remote: 'https://gitlab.com/example-project/-/raw/master/.gitlab-ci.yml'

All nested includes execute without context as a public user, so you can only include public projects or templates.

include:template

Introduced in GitLab 11.7.

Use include:template to include .gitlab-ci.yml templates that are shipped with GitLab.

For example:

# File sourced from the GitLab template collection
include:
  - template: Auto-DevOps.gitlab-ci.yml

Multiple include:template files:

include:
  - template: Android-Fastlane.gitlab-ci.yml
  - template: Auto-DevOps.gitlab-ci.yml

All nested includes are executed only with the permission of the user, so it’s possible to use project, remote or template includes.

Nested includes

Introduced in GitLab 11.9.

Use nested includes to compose a set of includes.

You can have up to 100 includes, but you can’t have duplicate includes.

In GitLab 12.4 and later, the time limit to resolve all files is 30 seconds.

Additional includes examples

View additional includes examples.

Keyword details

The following topics explain how to use keywords to configure CI/CD pipelines.

image

Use image to specify a Docker image to use for the job.

For:

• Usage examples, see Define image and services from .gitlab-ci.yml.
• Detailed usage information, refer to Docker integration documentation.

image:name

An extended Docker configuration option.

For more information, see Available settings for image.

image:entrypoint

An extended Docker configuration option.

For more information, see Available settings for image.

services

Use services to specify a service Docker image, linked to a base image specified in image.

For:

• Usage examples, see Define image and services from .gitlab-ci.yml.
• Detailed usage information, refer to Docker integration documentation.
• Example services, see GitLab CI/CD Services.

services:name

An extended Docker configuration option.

For more information, see Available settings for services.

services:alias

An extended Docker configuration option.

For more information, see Available settings for services.

services:entrypoint

An extended Docker configuration option.

For more information, see Available settings for services.

services:command

An extended Docker configuration option.

For more information, see Available settings for services.

script

Use script to specify a shell script for the runner to execute.

All jobs except trigger jobs require a script keyword.

For example:

job:
  script: "bundle exec rspec"

You can use YAML anchors with script.

The script keyword can also contain several commands in an array:

job:
  script:
    - uname -a
    - bundle exec rspec

Sometimes, script commands must be wrapped in single or double quotes. For example, commands that contain a colon (:) must be wrapped in single quotes ('). The YAML parser needs to interpret the text as a string rather than a “key: value” pair.

For example, this script uses a colon:

job:
  script:
    - curl --request POST --header 'Content-Type: application/json' "https://gitlab/api/v4/projects"

To be considered valid YAML, you must wrap the entire command in single quotes. If the command already uses single quotes, you should change them to double quotes (") if possible:

job:
  script:
    - 'curl --request POST --header "Content-Type: application/json" "https://gitlab/api/v4/projects"'

You can verify the syntax is valid with the CI Lint tool.

Be careful when using these characters as well:

• {, }, [, ], ,, &, *, #, ?, |, -, <, >, =, !, %, @, `.

If any of the script commands return an exit code other than zero, the job fails and further commands are not executed. Store the exit code in a variable to avoid this behavior:

job:
  script:
    - false || exit_code=$?
    - if [ $exit_code -ne 0 ]; then echo "Previous command failed"; fi;

before_script

Use before_script to define an array of commands that should run before each job, but after artifacts are restored.

Scripts you specify in before_script are concatenated with any scripts you specify in the main script. The combine scripts execute together in a single shell.

You can overwrite a globally-defined before_script if you define it in a job:

default:
  before_script:
    - echo "Execute this script in all jobs that don't already have a before_script section."

job1:
  script:
    - echo "This script executes after the global before_script."

job:
  before_script:
    - echo "Execute this script instead of the global before_script."
  script:
    - echo "This script executes after the job's `before_script`"

You can use YAML anchors with before_script.

after_script

Use after_script to define an array of commands that run after each job, including failed jobs.

If a job times out or is cancelled, the after_script commands do not execute. An issue exists to support executing after_script commands for timed-out or cancelled jobs.

Scripts you specify in after_script execute in a new shell, separate from any before_script or script scripts. As a result, they:

• Have a current working directory set back to the default.
• Have no access to changes done by scripts defined in before_script or script, including:
  • Command aliases and variables exported in script scripts.
  • Changes outside of the working tree (depending on the runner executor), like software installed by a before_script or script script.
• Have a separate timeout, which is hard coded to 5 minutes. See the related issue for details.
• Don’t affect the job’s exit code. If the script section succeeds and the after_script times out or fails, the job exits with code 0 (Job Succeeded).

default:
  after_script:
    - echo "Execute this script in all jobs that don't already have an after_script section."

job1:
  script:
    - echo "This script executes first. When it completes, the global after_script executes."

job:
  script:
    - echo "This script executes first. When it completes, the job's `after_script` executes."
  after_script:
    - echo "Execute this script instead of the global after_script."

You can use YAML anchors with after_script.

Script syntax

You can use syntax in script sections to:

• Split long commands into multiline commands.
• Use color codes to make job logs easier to review.
• Create custom collapsible sections to simplify job log output.

stage

Use stage to define which stage a job runs in. Jobs in the same stage can execute in parallel (subject to certain conditions).

Jobs without a stage entry use the test stage by default. If you do not define stages in the pipeline, you can use the 5 default stages, which execute in this order:

• .pre
• build
• test
• deploy
• .post For example:

stages:
  - build
  - test
  - deploy

job 0:
  stage: .pre
  script: make something useful before build stage

job 1:
  stage: build
  script: make build dependencies

job 2:
  stage: build
  script: make build artifacts

job 3:
  stage: test
  script: make test

job 4:
  stage: deploy
  script: make deploy

job 5:
  stage: .post
  script: make something useful at the end of pipeline

Use your own runners

When you use your own runners, each runner runs only one job at a time by default. Jobs can run in parallel if they run on different runners.

If you have only one runner, jobs can run in parallel if the runner’s concurrent setting is greater than 1.

.pre and .post

Introduced in GitLab 12.4.

Use pre and post for jobs that need to run first or last in a pipeline.

• .pre is guaranteed to always be the first stage in a pipeline.
• .post is guaranteed to always be the last stage in a pipeline.

User-defined stages are executed after .pre and before .post.

You must have a job in at least one stage other than .pre or .post.

You can’t change the order of .pre and .post, even if you define them out of order in the .gitlab-ci.yml file. For example, the following configurations are equivalent:

stages:
  - .pre
  - a
  - b
  - .post

stages:
  - a
  - .pre
  - b
  - .post

stages:
  - a
  - b

extends

Introduced in GitLab 11.3.

Use extends to reuse configuration sections. It’s an alternative to YAML anchors and is a little more flexible and readable. You can use extends to reuse configuration from included configuration files.

In the following example, the rspec job uses the configuration from the .tests template job. GitLab:

• Performs a reverse deep merge based on the keys.
• Merges the .tests content with the rspec job.
• Doesn’t merge the values of the keys.

.tests:
  script: rake test
  stage: test
  only:
    refs:
      - branches

rspec:
  extends: .tests
  script: rake rspec
  only:
    variables:
      - $RSPEC

The result is this rspec job:

rspec:
  script: rake rspec
  stage: test
  only:
    refs:
      - branches
    variables:
      - $RSPEC

.tests in this example is a hidden job, but it’s possible to extend configuration from regular jobs as well.

extends supports multi-level inheritance. You should avoid using more than three levels, but you can use as many as eleven. The following example has two levels of inheritance:

.tests:
  only:
    - pushes

.rspec:
  extends: .tests
  script: rake rspec

rspec 1:
  variables:
    RSPEC_SUITE: '1'
  extends: .rspec

rspec 2:
  variables:
    RSPEC_SUITE: '2'
  extends: .rspec

spinach:
  extends: .tests
  script: rake spinach

In GitLab 12.0 and later, it’s also possible to use multiple parents for extends.

Merge details

You can use extends to merge hashes but not arrays. The algorithm used for merge is “closest scope wins,” so keys from the last member always override anything defined on other levels. For example:

.only-important:
  variables:
    URL: "http://my-url.internal"
    IMPORTANT_VAR: "the details"
  only:
    - master
    - stable
  tags:
    - production
  script:
    - echo "Hello world!"

.in-docker:
  variables:
    URL: "http://docker-url.internal"
  tags:
    - docker
  image: alpine

rspec:
  variables:
    GITLAB: "is-awesome"
  extends:
    - .only-important
    - .in-docker
  script:
    - rake rspec

The result is this rspec job:

rspec:
  variables:
    URL: "http://docker-url.internal"
    IMPORTANT_VAR: "the details"
    GITLAB: "is-awesome"
  only:
    - master
    - stable
  tags:
    - docker
  image: alpine
  script:
    - rake rspec

In this example:

• The variables sections merge, but URL: "http://docker-url.internal" overwrites URL: "http://my-url.internal".
• tags: ['docker'] overwrites tags: ['production'].
• script does not merge, but script: ['rake rspec'] overwrites script: ['echo "Hello world!"']. You can use YAML anchors to merge arrays.

Use extends and include together

To reuse configuration from different configuration files, combine extends and include.

In the following example, a script is defined in the included.yml file. Then, in the .gitlab-ci.yml file, extends refers to the contents of the script:

rules

Introduced in GitLab 12.3.

Use rules to include or exclude jobs in pipelines.

Rules are evaluated in order until the first match. When a match is found, the job is either included or excluded from the pipeline, depending on the configuration. The job can also have certain attributes added to it.

rules replaces only/except and they can’t be used together in the same job. If you configure one job to use both keywords, the linter returns a key may not be used with rules error.

Rules attributes

The job attributes you can use with rules are:

• when: If not defined, defaults to when: on_success.
  • If used as when: delayed, start_in is also required.
• allow_failure: If not defined, defaults to allow_failure: false.
• variables: If not defined, uses the variables defined elsewhere.

If a rule evaluates to true, and when has any value except never, the job is included in the pipeline.

For example:

docker build:
  script: docker build -t my-image:$CI_COMMIT_REF_SLUG .
  rules:
    - if: '$CI_COMMIT_BRANCH == "master"'
      when: delayed
      start_in: '3 hours'
      allow_failure: true

Rules clauses

Available rule clauses are:

Rules are evaluated in order until a match is found. If a match is found, the attributes are checked to see if the job should be added to the pipeline. If no attributes are defined, the defaults are:

• when: on_success
• allow_failure: false

The job is added to the pipeline:

• If a rule matches and has when: on_success, when: delayed or when: always.
• If no rules match, but the last clause is when: on_success, when: delayed or when: always (with no rule).

The job is not added to the pipeline:

• If no rules match, and there is no standalone when: on_success, when: delayed or when: always.
• If a rule matches, and has when: never as the attribute.

The following example uses if to strictly limit when jobs run:

job:
  script: echo "Hello, Rules!"
  rules:
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event"'
      when: manual
      allow_failure: true
    - if: '$CI_PIPELINE_SOURCE == "schedule"'

• If the pipeline is for a merge request, the first rule matches, and the job is added to the merge request pipeline with attributes of:
  • when: manual (manual job)
  • allow_failure: true (the pipeline continues running even if the manual job is not run)
• If the pipeline is not for a merge request, the first rule doesn’t match, and the second rule is evaluated.
• If the pipeline is a scheduled pipeline, the second rule matches, and the job is added to the scheduled pipeline. No attributes were defined, so it is added with:
  • when: on_success (default)
  • allow_failure: false (default)
• In all other cases, no rules match, so the job is not added to any other pipeline.

Alternatively, you can define a set of rules to exclude jobs in a few cases, but run them in all other cases:

job:
  script: echo "Hello, Rules!"
  rules:
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event"'
      when: never
    - if: '$CI_PIPELINE_SOURCE == "schedule"'
      when: never
    - when: on_success

• If the pipeline is for a merge request, the job is not added to the pipeline.
• If the pipeline is a scheduled pipeline, the job is not added to the pipeline.
• In all other cases, the job is added to the pipeline, with when: on_success.

WARNING: If you use a when: clause as the final rule (not including when: never), two simultaneous pipelines may start. Both push pipelines and merge request pipelines can be triggered by the same event (a push to the source branch for an open merge request). See how to prevent duplicate pipelines for more details.

Avoid duplicate pipelines

If a job uses rules, a single action, like pushing a commit to a branch, can trigger multiple pipelines. You don’t have to explicitly configure rules for multiple types of pipeline to trigger them accidentally.

Some configurations that have the potential to cause duplicate pipelines cause a pipeline warning to be displayed. Introduced in GitLab 13.3.

For example:

job:
  script: echo "This job creates double pipelines!"
  rules:
    - if: '$CUSTOM_VARIABLE == "false"'
      when: never
    - when: always

This job does not run when $CUSTOM_VARIABLE is false, but it does run in all other pipelines, including both push (branch) and merge request pipelines. With this configuration, every push to an open merge request’s source branch causes duplicated pipelines.

To avoid duplicate pipelines, you can:

You can also avoid duplicate pipelines by changing the job rules to avoid either push (branch) pipelines or merge request pipelines. However, if you use a - when: always rule without workflow: rules, GitLab still displays a pipeline warning.

For example, the following does not trigger double pipelines, but is not recommended without workflow: rules:

job:
  script: echo "This job does NOT create double pipelines!"
  rules:
    - if: '$CI_PIPELINE_SOURCE == "push"'
      when: never
    - when: always

You should not include both push and merge request pipelines in the same job without workflow:rules that prevent duplicate pipelines:

job:
  script: echo "This job creates double pipelines!"
  rules:
    - if: '$CI_PIPELINE_SOURCE == "push"'
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event"'

Also, do not mix only/except jobs with rules jobs in the same pipeline. It may not cause YAML errors, but the different default behaviors of only/except and rules can cause issues that are difficult to troubleshoot:

job-with-no-rules:
  script: echo "This job runs in branch pipelines."

job-with-rules:
  script: echo "This job runs in merge request pipelines."
  rules:
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event"'

For every change pushed to the branch, duplicate pipelines run. One branch pipeline runs a single job (job-with-no-rules), and one merge request pipeline runs the other job (job-with-rules). Jobs with no rules default to except: merge_requests, so job-with-no-rules runs in all cases except merge requests.

rules:if

Use rules:if clauses to specify when to add a job to a pipeline:

• If an if statement is true, add the job to the pipeline.
• If an if statement is true, but it’s combined with when: never, do not add the job to the pipeline.
• If no if statements are true, do not add the job to the pipeline.

rules:if differs slightly from only:variables by accepting only a single expression string per rule, rather than an array of them. Any set of expressions to be evaluated can be conjoined into a single expression by using && or ||, and the variable matching operators (==, !=, =~ and !~).

Unlike variables in script sections, variables in rules expressions are always formatted as $VARIABLE.

if: clauses are evaluated based on the values of predefined CI/CD variables or custom CI/CD variables.

For example:

job:
  script: echo "Hello, Rules!"
  rules:
    - if: '$CI_MERGE_REQUEST_SOURCE_BRANCH_NAME =~ /^feature/ && $CI_MERGE_REQUEST_TARGET_BRANCH_NAME == "master"'
      when: always
    - if: '$CI_MERGE_REQUEST_SOURCE_BRANCH_NAME =~ /^feature/'
      when: manual
      allow_failure: true
    - if: '$CI_MERGE_REQUEST_SOURCE_BRANCH_NAME'  # Checking for the presence of a variable is possible

Some details regarding the logic that determines the when for the job:

• If none of the provided rules match, the job is set to when: never and is not included in the pipeline.
• A rule without any conditional clause, such as a when or allow_failure rule without if or changes, always matches, and is always used if reached.
• If a rule matches and has no when defined, the rule uses the when defined for the job, which defaults to on_success if not defined.
• You can define when once per rule, or once at the job-level, which applies to all rules. You can’t mix when at the job-level with when in rules.

Common if clauses for rules

For behavior similar to the only/except keywords, you can check the value of the $CI_PIPELINE_SOURCE variable:

The following example runs the job as a manual job in scheduled pipelines or in push pipelines (to branches or tags), with when: on_success (default). It does not add the job to any other pipeline type.

job:
  script: echo "Hello, Rules!"
  rules:
    - if: '$CI_PIPELINE_SOURCE == "schedule"'
      when: manual
      allow_failure: true
    - if: '$CI_PIPELINE_SOURCE == "push"'

The following example runs the job as a when: on_success job in merge request pipelines and scheduled pipelines. It does not run in any other pipeline type.

job:
  script: echo "Hello, Rules!"
  rules:
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event"'
    - if: '$CI_PIPELINE_SOURCE == "schedule"'

Other commonly used variables for if clauses:

• if: $CI_COMMIT_TAG: If changes are pushed for a tag.
• if: $CI_COMMIT_BRANCH: If changes are pushed to any branch.
• if: '$CI_COMMIT_BRANCH == "main"': If changes are pushed to main.
• if: 'CI_COMMIT_BRANCH == CI_DEFAULT_BRANCH': If changes are pushed to the default branch. Use when you want to have the same configuration in multiple projects with different default branches.
• if: '$CI_COMMIT_BRANCH =~ /regex-expression/': If the commit branch matches a regular expression.
• if: '$CUSTOM_VARIABLE !~ /regex-expression/': If the custom variable CUSTOM_VARIABLE does not match a regular expression.
• if: '$CUSTOM_VARIABLE == "value1"': If the custom variable CUSTOM_VARIABLE is exactly value1.

rules:changes

Use rules:changes to specify when to add a job to a pipeline by checking for changes to specific files.

rules: changes works the same way as only: changes and except: changes. It accepts an array of paths. You should use rules: changes only with branch pipelines or merge request pipelines. For example, it’s common to use rules: changes with merge request pipelines:

docker build:
  script: docker build -t my-image:$CI_COMMIT_REF_SLUG .
  rules:
    - if: '$CI_PIPELINE_SOURCE == "merge_request_event"'
      changes:
        - Dockerfile
      when: manual
      allow_failure: true

In this example:

• If the pipeline is a merge request pipeline, check Dockerfile for changes.
• If Dockerfile has changed, add the job to the pipeline as a manual job, and the pipeline continues running even if the job is not triggered (allow_failure: true).
• If Dockerfile has not changed, do not add job to any pipeline (same as when: never).

To use rules: changes with branch pipelines instead of merge request pipelines, change the if: clause in the previous example to:

rules:
  - if: $CI_PIPELINE_SOURCE == "push" && $CI_COMMIT_BRANCH

To implement a rule similar to except:changes, use when: never.

WARNING: You can use rules: changes with other pipeline types, but it is not recommended because rules: changes always evaluates to true when there is no Git push event. Tag pipelines, scheduled pipelines, and so on do not have a Git push event associated with them. A rules: changes job is always added to those pipeline if there is no if: statement that limits the job to branch or merge request pipelines.

Variables in rules:changes

• Introduced in GitLab 13.6.
• Feature flag removed in GitLab 13.7.

You can use CI/CD variables in rules:changes expressions to determine when to add jobs to a pipeline:

docker build:
  variables:
    DOCKERFILES_DIR: 'path/to/files/'
  script: docker build -t my-image:$CI_COMMIT_REF_SLUG .
  rules:
    - changes:
        - $DOCKERFILES_DIR/*

You can use the character for both variables and paths. For example, if the DOCKERFILES_DIR variable exists, its value is used. If it does not exist, the

rules:exists

Introduced in GitLab 12.4.

Use exists to run a job when certain files exist in the repository. You can use an array of paths.

In the following example, job runs if a Dockerfile exists anywhere in the repository:

job:
  script: docker build -t my-image:$CI_COMMIT_REF_SLUG .
  rules:
    - exists:
        - Dockerfile

Paths are relative to the project directory ($CI_PROJECT_DIR) and can’t directly link outside it.

You can use glob patterns to match multiple files in any directory in the repository:

job:
  script: bundle exec rspec
  rules:
    - exists:
        - spec/**.rb

Glob patterns are interpreted with Ruby File.fnmatch with the flags File::FNM_PATHNAME | File::FNM_DOTMATCH | File::FNM_EXTGLOB.

For performance reasons, GitLab matches a maximum of 10,000 exists patterns. After the 10,000th check, rules with patterned globs always match.

rules:allow_failure

Introduced in GitLab 12.8.

You can use allow_failure: true in rules: to allow a job to fail, or a manual job to wait for action, without stopping the pipeline itself. All jobs that use rules: default to allow_failure: false if you do not define allow_failure:.

The rule-level rules:allow_failure option overrides the job-level allow_failure option, and is only applied when the particular rule triggers the job.

job:
  script: echo "Hello, Rules!"
  rules:
    - if: '$CI_MERGE_REQUEST_TARGET_BRANCH_NAME == "master"'
      when: manual
      allow_failure: true

In this example, if the first rule matches, then the job has when: manual and allow_failure: true.

rules:variables

• Introduced in GitLab 13.7.
• Feature flag removed in GitLab 13.10.

Use variables in rules: to define variables for specific conditions.

For example:

job:
  variables:
    DEPLOY_VARIABLE: "default-deploy"
  rules:
    - if: $CI_COMMIT_REF_NAME =~ /master/
      variables:                              # Override DEPLOY_VARIABLE defined
        DEPLOY_VARIABLE: "deploy-production"  # at the job level.
    - if: $CI_COMMIT_REF_NAME =~ /feature/
      variables:
        IS_A_FEATURE: "true"                  # Define a new variable.
  script:
    - echo "Run script with $DEPLOY_VARIABLE as an argument"
    - echo "Run another script if $IS_A_FEATURE exists"

Complex rule clauses

To conjoin if, changes, and exists clauses with an AND, use them in the same rule.

In the following example:

• If the Dockerfile file or any file in /docker/scripts has changed, and $VAR == “string value”, then the job runs manually
• Otherwise, the job isn’t included in the pipeline.

docker build:
  script: docker build -t my-image:$CI_COMMIT_REF_SLUG .
  rules:
    - if: '$VAR == "string value"'
      changes:  # Include the job and set to when:manual if any of the follow paths match a modified file.
        - Dockerfile
        - docker/scripts/*
      when: manual
      # - "when: never" would be redundant here. It is implied any time rules are listed.

Keywords such as branches or refs that are available for only/except are not available in rules. They are being individually considered for their usage and behavior in this context. Future keyword improvements are being discussed in our epic for improving rules, where anyone can add suggestions or requests.

You can use parentheses with && and || to build more complicated variable expressions. Introduced in GitLab 13.3:

job1:
  script:
    - echo This rule uses parentheses.
  rules:
    if: ($CI_COMMIT_BRANCH == "master" || $CI_COMMIT_BRANCH == "develop") && $MY_VARIABLE

WARNING: Before GitLab 13.3, rules that use both || and && may evaluate with an unexpected order of operations.

only/except (basic)

NOTE: only and except are not being actively developed. To define when to add jobs to pipelines, use rules.

only and except are two keywords that determine when to add jobs to pipelines:

1. only defines the names of branches and tags the job runs for.
2. except defines the names of branches and tags the job does not run for.

A few rules apply to the usage of job policy:

• only and except are inclusive. If both only and except are defined in a job specification, the ref is filtered by only and except.
• only and except can use regular expressions (supported regexp syntax).
• only and except can specify a repository path to filter jobs for forks.

In addition, only and except can use these keywords:

Scheduled pipelines run on specific branches, so jobs configured with only: branches run on scheduled pipelines too. Add except: schedules to prevent jobs with only: branches from running on scheduled pipelines.

In the following example, job runs only for refs that start with issue-. All branches are skipped:

job:
  # use regexp
  only:
    - /^issue-.*$/
  # use special keyword
  except:
    - branches

Pattern matching is case-sensitive by default. Use the i flag modifier, like /pattern/i, to make a pattern case-insensitive:

job:
  # use regexp
  only:
    - /^issue-.*$/i
  # use special keyword
  except:
    - branches

In the following example, job runs only for:

• Git tags
• Triggers
• Scheduled pipelines

job:
  # use special keywords
  only:
    - tags
    - triggers
    - schedules

To execute jobs only for the parent repository and not forks:

job:
  only:
    - branches@gitlab-org/gitlab
  except:
    - master@gitlab-org/gitlab
    - /^release/.*$/@gitlab-org/gitlab

This example runs job for all branches on gitlab-org/gitlab, except master and branches that start with release/.

If a job does not have an only rule, only: ['branches', 'tags'] is set by default. If the job does not have an except rule, it’s empty.

For example, job1 and job2 are essentially the same:

job1:
  script: echo 'test'

job2:
  script: echo 'test'
  only: ['branches', 'tags']

Regular expressions

The @ symbol denotes the beginning of a ref’s repository path. To match a ref name that contains the @ character in a regular expression, you must use the hex character code match \x40.

Only the tag or branch name can be matched by a regular expression. The repository path, if given, is always matched literally.

To match the tag or branch name, the entire ref name part of the pattern must be a regular expression surrounded by /. For example, you can’t use issue-/.*/ to match all tag names or branch names that begin with issue-, but you can use /issue-.*/.

Regular expression flags must be appended after the closing /.

NOTE: Use anchors ^ and to avoid the regular expression matching only a substring of the tag name or branch name. For example, /^issue-.*/ is equivalent to /^issue-/, while just /issue/ would also match a branch called severe-issues.

Supported only/except regexp syntax

In GitLab 11.9.4, GitLab began internally converting the regexp used in only and except keywords to RE2.

RE2 limits the set of available features due to computational complexity, and some features, like negative lookaheads, became unavailable. Only a subset of features provided by Ruby Regexp are now supported.

From GitLab 11.9.7 to GitLab 12.0, GitLab provided a feature flag to let you use unsafe regexp syntax. After migrating to safe syntax, you should disable this feature flag again:

Feature.enable(:allow_unsafe_ruby_regexp)

only/except (advanced)

GitLab supports multiple strategies, and it’s possible to use an array or a hash configuration scheme.

Four keys are available:

• refs
• variables
• changes
• kubernetes

If you use multiple keys under only or except, the keys are evaluated as a single conjoined expression. That is:

• only: includes the job if all of the keys have at least one condition that matches.
• except: excludes the job if any of the keys have at least one condition that matches.

With only, individual keys are logically joined by an AND. A job is added to the pipeline if the following is true:

• (any listed refs are true) AND (any listed variables are true) AND (any listed changes are true) AND (any chosen Kubernetes status matches)

In the following example, the test job is only created when all of the following are true:

• The pipeline is scheduled or runs for master.
• The variables keyword matches.
• The kubernetes service is active on the project.

test:
  script: npm run test
  only:
    refs:
      - master
      - schedules
    variables:
      - $CI_COMMIT_MESSAGE =~ /run-end-to-end-tests/
    kubernetes: active

With except, individual keys are logically joined by an OR. A job is not added if the following is true:

• (any listed refs are true) OR (any listed variables are true) OR (any listed changes are true) OR (a chosen Kubernetes status matches)

In the following example, the test job is not created when any of the following are true:

• The pipeline runs for the master branch.
• There are changes to the README.md file in the root directory of the repository.

test:
  script: npm run test
  except:
    refs:
      - master
    changes:
      - "README.md"

only:refs/except:refs

refs policy introduced in GitLab 10.0.

The refs strategy can take the same values as the simplified only/except configuration.

In the following example, the deploy job is created only when the pipeline is scheduled or runs for the master branch:

deploy:
  only:
    refs:
      - master
      - schedules

only:kubernetes/except:kubernetes

kubernetes policy introduced in GitLab 10.0.

The kubernetes strategy accepts only the active keyword.

In the following example, the deploy job is created only when the Kubernetes service is active in the project:

deploy:
  only:
    kubernetes: active

only:variables/except:variables

variables policy introduced in GitLab 10.7.

The variables keyword defines variable expressions.

These expressions determine whether or not a job should be created.

Examples of variable expressions:

deploy:
  script: cap staging deploy
  only:
    refs:
      - branches
    variables:
      - $RELEASE == "staging"
      - $STAGING

Another use case is excluding jobs depending on a commit message:

end-to-end:
  script: rake test:end-to-end
  except:
    variables:
      - $CI_COMMIT_MESSAGE =~ /skip-end-to-end-tests/

You can use parentheses with && and || to build more complicated variable expressions. Introduced in GitLab 13.3:

job1:
  script:
    - echo This rule uses parentheses.
  only:
    variables:
      - ($CI_COMMIT_BRANCH == "master" || $CI_COMMIT_BRANCH == "develop") && $MY_VARIABLE

only:changes/except:changes

changes policy introduced in GitLab 11.4.

Use the changes keyword with only to run a job, or with except to skip a job, when a Git push event modifies a file.

Use only:changes with pipelines triggered by the following refs only:

• branches
• external_pull_requests
• merge_requests (see additional details about using only:changes with pipelines for merge requests)

WARNING: In pipelines with sources other than the three above changes can’t determine if a given file is new or old and always returns true. You can configure jobs to use only: changes with other only: refs keywords. However, those jobs ignore the changes and always run.

In the following example, when you push commits to an existing branch, the docker build job runs only if any of these files change:

• The Dockerfile file.
• Files in the docker/scripts/ directory.
• Files and subdirectories in the dockerfiles directory.
• Files with rb, py, sh extensions in the more_scripts directory.

docker build:
  script: docker build -t my-image:$CI_COMMIT_REF_SLUG .
  only:
    refs:
      - branches
    changes:
      - Dockerfile
      - docker/scripts/*
      - dockerfiles/**/*
      - more_scripts/*.{rb,py,sh}

WARNING: If you use only:changes with only allow merge requests to be merged if the pipeline succeeds, you should also use only:merge_requests. Otherwise it may not work as expected.

You can also use glob patterns to match multiple files in either the root directory of the repository, or in any directory in the repository. However, they must be wrapped in double quotes or GitLab can’t parse them:

test:
  script: npm run test
  only:
    refs:
      - branches
    changes:
      - "*.json"
      - "**/*.sql"

You can skip a job if a change is detected in any file with a .md extension in the root directory of the repository:

build:
  script: npm run build
  except:
    changes:
      - "*.md"

If you change multiple files, but only one file ends in .md, the build job is still skipped. The job does not run for any of the files.

Read more about how to use this feature with:

• New branches or tags without pipelines for merge requests.
• Scheduled pipelines.

Use only:changes with pipelines for merge requests

With pipelines for merge requests, it’s possible to define a job to be created based on files modified in a merge request.

Use this keyword with only: [merge_requests] so GitLab can find the correct base SHA of the source branch. File differences are correctly calculated from any further commits, and all changes in the merge requests are properly tested in pipelines.

For example:

docker build service one:
  script: docker build -t my-service-one-image:$CI_COMMIT_REF_SLUG .
  only:
    refs:
      - merge_requests
    changes:
      - Dockerfile
      - service-one/**/*

In this scenario, if a merge request changes files in the service-one directory or the Dockerfile, GitLab creates the docker build service one job.

For example:

docker build service one:
  script: docker build -t my-service-one-image:$CI_COMMIT_REF_SLUG .
  only:
    changes:
      - Dockerfile
      - service-one/**/*

In this example, the pipeline might fail because of changes to a file in service-one/**/*.

A later commit that doesn’t have changes in service-one/**/* but does have changes to the Dockerfile can pass. The job only tests the changes to the Dockerfile.

GitLab checks the most recent pipeline that passed. If the merge request is mergeable, it doesn’t matter that an earlier pipeline failed because of a change that has not been corrected.

When you use this configuration, ensure that the most recent pipeline properly corrects any failures from previous pipelines.

Use only:changes without pipelines for merge requests

Without pipelines for merge requests, pipelines run on branches or tags that don’t have an explicit association with a merge request. In this case, a previous SHA is used to calculate the diff, which is equivalent to git diff HEAD~. This can result in some unexpected behavior, including:

• When pushing a new branch or a new tag to GitLab, the policy always evaluates to true.
• When pushing a new commit, the changed files are calculated by using the previous commit as the base SHA.

Use only:changes with scheduled pipelines

only:changes always evaluates as true in Scheduled pipelines. All files are considered to have changed when a scheduled pipeline runs.

needs

• Introduced in GitLab 12.2.
• In GitLab 12.3, maximum number of jobs in needs array raised from five to 50.
• Introduced in GitLab 12.8, needs: [] lets jobs start immediately.

Use needs: to execute jobs out-of-order. Relationships between jobs that use needs can be visualized as a directed acyclic graph.

You can ignore stage ordering and run some jobs without waiting for others to complete. Jobs in multiple stages can run concurrently.

The following example creates four paths of execution:

• Linter: the lint job runs immediately without waiting for the build stage to complete because it has no needs (needs: []).
• Linux path: the linux:rspec and linux:rubocop jobs runs as soon as the linux:build job finishes without waiting for mac:build to finish.
• macOS path: the mac:rspec and mac:rubocop jobs runs as soon as the mac:build job finishes, without waiting for linux:build to finish.
• The production job runs as soon as all previous jobs finish; in this case: linux:build, linux:rspec, linux:rubocop, mac:build, mac:rspec, mac:rubocop.

linux:build:
  stage: build

mac:build:
  stage: build

lint:
  stage: test
  needs: []

linux:rspec:
  stage: test
  needs: ["linux:build"]

linux:rubocop:
  stage: test
  needs: ["linux:build"]

mac:rspec:
  stage: test
  needs: ["mac:build"]

mac:rubocop:
  stage: test
  needs: ["mac:build"]

production:
  stage: deploy

Requirements and limitations

• In GitLab 13.9 and older, if needs: refers to a job that might not be added to a pipeline because of only, except, or rules, the pipeline might fail to create.
• The maximum number of jobs that a single job can need in the needs: array is limited:
  • For GitLab.com, the limit is 50. For more information, see our infrastructure issue.
  • For self-managed instances, the limit is: 50. This limit can be changed.
• If needs: refers to a job that uses the parallel keyword, it depends on all jobs created in parallel, not just one job. It also downloads artifacts from all the parallel jobs by default. If the artifacts have the same name, they overwrite each other and only the last one downloaded is saved.
• needs: is similar to dependencies: in that it must use jobs from prior stages, meaning it’s impossible to create circular dependencies. Depending on jobs in the current stage is not possible either, but support is planned.
• Stages must be explicitly defined for all jobs that have the keyword needs: or are referred to by one.

Changing the needs: job limit (FREE SELF)

The maximum number of jobs that can be defined in needs: defaults to 50.

A GitLab administrator with access to the GitLab Rails console can choose a custom limit. For example, to set the limit to 100:

Plan.default.actual_limits.update!(ci_needs_size_limit: 100)

To disable directed acyclic graphs (DAG), set the limit to 0.

Artifact downloads with needs

Introduced in GitLab v12.6.

When a job uses needs, it no longer downloads all artifacts from previous stages by default, because jobs with needs can start before earlier stages complete. With needs you can only download artifacts from the jobs listed in the needs: configuration.

Use artifacts: true (default) or artifacts: false to control when artifacts are downloaded in jobs that use needs.

In the following example, the rspec job downloads the build_job artifacts, but the rubocop job does not:

build_job:
  stage: build
  artifacts:
    paths:
      - binaries/

rspec:
  stage: test
  needs:
    - job: build_job
      artifacts: true

rubocop:
  stage: test
  needs:
    - job: build_job
      artifacts: false

In the following example, the rspec job downloads the artifacts from all three build_jobs. artifacts is:

• Set to true for build_job_1.
• Defaults to true for both build_job_2 and build_job_3.

rspec:
  needs:
    - job: build_job_1
      artifacts: true
    - job: build_job_2
    - build_job_3

In GitLab 12.6 and later, you can’t combine the dependencies keyword with needs.

Cross project artifact downloads with needs (PREMIUM)

Introduced in GitLab v12.7.

Use needs to download artifacts from up to five jobs in pipelines:

• On other refs in the same project.
• In different projects, groups and namespaces.

build_job:
  stage: build
  script:
    - ls -lhR
  needs:
    - project: namespace/group/project-name
      job: build-1
      ref: master
      artifacts: true

build_job downloads the artifacts from the latest successful build-1 job on the master branch in the group/project-name project. If the project is in the same group or namespace, you can omit them from the project: keyword. For example, project: group/project-name or project: project-name.

The user running the pipeline must have at least reporter access to the group or project, or the group/project must have public visibility.

Artifact downloads between pipelines in the same project

Use needs to download artifacts from different pipelines in the current project. Set the project keyword as the current project’s name, and specify a ref.

In the following example, build_job downloads the artifacts for the latest successful build-1 job with the other-ref ref:

build_job:
  stage: build
  script:
    - ls -lhR
  needs:
    - project: group/same-project-name
      job: build-1
      ref: other-ref
      artifacts: true

CI/CD variable support for project:, job:, and ref was introduced in GitLab 13.3. Feature flag removed in GitLab 13.4.

For example:

build_job:
  stage: build
  script:
    - ls -lhR
  needs:
    - project: $CI_PROJECT_PATH
      job: $DEPENDENCY_JOB_NAME
      ref: $ARTIFACTS_DOWNLOAD_REF
      artifacts: true

You can’t download artifacts from jobs that run in parallel:.

To download artifacts between parent-child pipelines, use needs:pipeline.

You should not download artifacts from the same ref as a running pipeline. Concurrent pipelines running on the same ref could override the artifacts.

Artifact downloads to child pipelines

Introduced in GitLab v13.7.

A child pipeline can download artifacts from a job in its parent pipeline or another child pipeline in the same parent-child pipeline hierarchy.

For example, with the following parent pipeline that has a job that creates some artifacts:

create-artifact:
  stage: build
  script: echo 'sample artifact' > artifact.txt
  artifacts:
    paths: [artifact.txt]

child-pipeline:
  stage: test
  trigger:
    include: child.yml
    strategy: depend
  variables:
    PARENT_PIPELINE_ID: $CI_PIPELINE_ID

A job in the child pipeline can download artifacts from the create-artifact job in the parent pipeline:

use-artifact:
  script: cat artifact.txt
  needs:
    - pipeline: $PARENT_PIPELINE_ID
      job: create-artifact

The pipeline attribute accepts a pipeline ID and it must be a pipeline present in the same parent-child pipeline hierarchy of the given pipeline.

The pipeline attribute does not accept the current pipeline ID ($CI_PIPELINE_ID). To download artifacts from a job in the current pipeline, use the basic form of needs.

Optional needs

• Introduced in GitLab 13.10.
• It’s deployed behind a feature flag, disabled by default.
• It’s disabled on GitLab.com.
• It’s not recommended for production use.
• To use it in GitLab self-managed instances, ask a GitLab administrator to enable it. (FREE SELF)

WARNING: This feature might not be available to you. Check the version history note above for details.

To need a job that sometimes does not exist in the pipeline, add optional: true to the needs configuration. If not defined, optional: false is the default.

Jobs that use rules, only, or except, might not always exist in a pipeline. When the pipeline starts, it checks the needs relationships before running. Without optional: true, needs relationships that point to a job that does not exist stops the pipeline from starting and causes a pipeline error similar to:

• 'job1' job needs 'job2' job, but it was not added to the pipeline

In this example:

• When the branch is master, the build job exists in the pipeline, and the rspec job waits for it to complete before starting.
• When the branch is not master, the build job does not exist in the pipeline. The rspec job runs immediately (similar to needs: []) because its needs relationship to the build job is optional.

build:
  stage: build
  rules:
    - if: $CI_COMMIT_REF_NAME == "master"

rspec:
  stage: test
  needs:
    - job: build
      optional: true

Enable or disable optional needs (FREE SELF)

Optional needs is under development and not ready for production use. It is deployed behind a feature flag that is disabled by default. GitLab administrators with access to the GitLab Rails console can enable it.

To enable it:

Feature.enable(:ci_needs_optional)

To disable it:

Feature.disable(:ci_needs_optional)

tags

Use tags to select a specific runner from the list of all runners that are available for the project.

When you register a runner, you can specify the runner’s tags, for example ruby, postgres, development.

In the following example, the job is run by a runner that has both ruby and postgres tags defined.

job:
  tags:
    - ruby
    - postgres

You can use tags to run different jobs on different platforms. For example, if you have an OS X runner with tag osx and a Windows runner with tag windows, you can run a job on each platform:

windows job:
  stage:
    - build
  tags:
    - windows
  script:
    - echo Hello, %USERNAME%!

osx job:
  stage:
    - build
  tags:
    - osx
  script:
    - echo "Hello, $USER!"

allow_failure

Use allow_failure when you want to let a job fail without impacting the rest of the CI suite. The default value is false, except for manual jobs that use the when: manual syntax.

In jobs that use rules:, all jobs default to allow_failure: false, including when: manual jobs.

When allow_failure is set to true and the job fails, the job shows an orange warning in the UI. However, the logical flow of the pipeline considers the job a success/passed, and is not blocked.

Assuming all other jobs are successful, the job’s stage and its pipeline show the same orange warning. However, the associated commit is marked as “passed”, without warnings.

In the following example, job1 and job2 run in parallel. If job1 fails, it doesn’t stop the next stage from running, because it’s marked with allow_failure: true:

job1:
  stage: test
  script:
    - execute_script_that_will_fail
  allow_failure: true

job2:
  stage: test
  script:
    - execute_script_that_will_succeed

job3:
  stage: deploy
  script:
    - deploy_to_staging

allow_failure:exit_codes

• Introduced in GitLab 13.8.
• Feature flag removed in GitLab 13.9.

Use allow_failure:exit_codes to dynamically control if a job should be allowed to fail. You can list which exit codes are not considered failures. The job fails for any other exit code:

test_job_1:
  script:
    - echo "Run a script that results in exit code 1. This job fails."
    - exit 1
  allow_failure:
    exit_codes: 137

test_job_2:
  script:
    - echo "Run a script that results in exit code 137. This job is allowed to fail."
    - exit 137
  allow_failure:
    exit_codes:
      - 137
      - 255

when

Use when to implement jobs that run in case of failure or despite the failure.

The valid values of when are:

1. on_success (default) - Execute job only when all jobs in earlier stages succeed, or are considered successful because they have allow_failure: true.
2. on_failure - Execute job only when at least one job in an earlier stage fails.
3. always - Execute job regardless of the status of jobs in earlier stages.
4. manual - Execute job manually.
5. delayed - Delay the execution of a job for a specified duration. Added in GitLab 11.14.
6. never:
   • With job rules, don’t execute job.
   • With workflow:rules, don’t run pipeline.

In the following example, the script:

1. Executes cleanup_build_job only when build_job fails.
2. Always executes cleanup_job as the last step in pipeline regardless of success or failure.
3. Executes deploy_job when you run it manually in the GitLab UI.

stages:
  - build
  - cleanup_build
  - test
  - deploy
  - cleanup

build_job:
  stage: build
  script:
    - make build

cleanup_build_job:
  stage: cleanup_build
  script:
    - cleanup build when failed
  when: on_failure

test_job:
  stage: test
  script:
    - make test

deploy_job:
  stage: deploy
  script:
    - make deploy
  when: manual

cleanup_job:
  stage: cleanup
  script:
    - cleanup after jobs
  when: always

when:manual

A manual job is a type of job that is not executed automatically and must be explicitly started by a user. You might want to use manual jobs for things like deploying to production.

To make a job manual, add when: manual to its configuration.

When the pipeline starts, manual jobs display as skipped and do not run automatically. They can be started from the pipeline, job, environment, and deployment views.

Manual jobs can be either optional or blocking:

• Optional: Manual jobs have `allow_failure: true set by default and are considered optional. The status of an optional manual job does not contribute to the overall pipeline status. A pipeline can succeed even if all its manual jobs fail.
• Blocking: To make a blocking manual job, add allow_failure: false to its configuration. Blocking manual jobs stop further execution of the pipeline at the stage where the job is defined. To let the pipeline continue running, click {play} (play) on the blocking manual job. Merge requests in projects with merge when pipeline succeeds enabled can’t be merged with a blocked pipeline. Blocked pipelines show a status of blocked.

When you use rules:, allow_failure defaults to false, including for manual jobs.

To trigger a manual job, a user must have permission to merge to the assigned branch. You can use protected branches to more strictly protect manual deployments from being run by unauthorized users.

In GitLab 13.5 and later, you can use when:manual in the same job as trigger. In GitLab 13.4 and earlier, using them together causes the error jobs:#{job-name} when should be on_success, on_failure or always.

Protecting manual jobs (PREMIUM)

Use protected environments to define a list of users authorized to run a manual job. You can authorize only the users associated with a protected environment to trigger manual jobs, which can:

• More precisely limit who can deploy to an environment.
• Block a pipeline until an approved user “approves” it.

To protect a manual job:

You can use protected environments with blocking manual jobs to have a list of users allowed to approve later pipeline stages. Add allow_failure: false to the protected manual job and the pipeline’s next stages only run after the manual job is triggered by authorized users.

when:delayed

Introduced in GitLab 11.4.

Use when: delayed to execute scripts after a waiting period, or if you want to avoid jobs immediately entering the pending state.

You can set the period with start_in keyword. The value of start_in is an elapsed time in seconds, unless a unit is provided. start_in must be less than or equal to one week. Examples of valid values include:

• '5'
• 5 seconds
• 30 minutes
• 1 day
• 1 week

When a stage includes a delayed job, the pipeline doesn’t progress until the delayed job finishes. You can use this keyword to insert delays between different stages.

The timer of a delayed job starts immediately after the previous stage completes. Similar to other types of jobs, a delayed job’s timer doesn’t start unless the previous stage passes.

The following example creates a job named timed rollout 10% that is executed 30 minutes after the previous stage completes:

timed rollout 10%:
  stage: deploy
  script: echo 'Rolling out 10% ...'
  when: delayed
  start_in: 30 minutes

To stop the active timer of a delayed job, click the {time-out} (Unschedule) button. This job can no longer be scheduled to run automatically. You can, however, execute the job manually.

To start a delayed job immediately, click the Play button. Soon GitLab Runner picks up and starts the job.

environment

Use environment to define the environment that a job deploys to. For example:

deploy to production:
  stage: deploy
  script: git push production HEAD:master
  environment: production

You can assign a value to the environment keyword by using:

• Plain text, like production.
• Variables, including CI/CD variables, predefined, secure, or variables defined in the .gitlab-ci.yml file.

You can’t use variables defined in a script section.

If you specify an environment and no environment with that name exists, an environment is created.

environment:name

Set a name for an environment. For example:

deploy to production:
  stage: deploy
  script: git push production HEAD:master
  environment:
    name: production

Common environment names are qa, staging, and production, but you can use any name you want.

You can assign a value to the name keyword by using:

• Plain text, like staging.
• Variables, including CI/CD variables, predefined, secure, or variables defined in the .gitlab-ci.yml file.

You can’t use variables defined in a script section.

The environment name can contain:

• Letters
• Digits
• Spaces
• -
• _
• /
• $
• {
• }

environment:url

Set a URL for an environment. For example:

deploy to production:
  stage: deploy
  script: git push production HEAD:master
  environment:
    name: production
    url: https://prod.example.com

After the job completes, you can access the URL by using a button in the merge request, environment, or deployment pages.

You can assign a value to the url keyword by using:

• Plain text, like https://prod.example.com.
• Variables, including CI/CD variables, predefined, secure, or variables defined in the .gitlab-ci.yml file.

You can’t use variables defined in a script section.

environment:on_stop

Closing (stopping) environments can be achieved with the on_stop keyword defined under environment. It declares a different job that runs to close the environment.

Read the environment:action section for an example.

environment:action

Use the action keyword to specify jobs that prepare, start, or stop environments.

Take for instance:

review_app:
  stage: deploy
  script: make deploy-app
  environment:
    name: review/$CI_COMMIT_REF_NAME
    url: https://$CI_ENVIRONMENT_SLUG.example.com
    on_stop: stop_review_app

stop_review_app:
  stage: deploy
  variables:
    GIT_STRATEGY: none
  script: make delete-app
  when: manual
  environment:
    name: review/$CI_COMMIT_REF_NAME
    action: stop

In the above example, the review_app job deploys to the review environment. A new stop_review_app job is listed under on_stop. After the review_app job is finished, it triggers the stop_review_app job based on what is defined under when. In this case, it is set to manual, so it needs a manual action from the GitLab UI to run.

Also in the example, GIT_STRATEGY is set to none. If the stop_review_app job is automatically triggered, the runner won’t try to check out the code after the branch is deleted.

The example also overwrites global variables. If your stop environment job depends on global variables, use anchor variables when you set the GIT_STRATEGY to change the job without overriding the global variables.

The stop_review_app job is required to have the following keywords defined:

• when, defined at either:
  • The job level.
  • In a rules clause. If you use rules: and when: manual, you should also set allow_failure: true so the pipeline can complete even if the job doesn’t run.
• environment:name
• environment:action

Additionally, both jobs should have matching rules or only/except configuration.

In the examples above, if the configuration is not identical:

• The stop_review_app job might not be included in all pipelines that include the review_app job.
• It is not possible to trigger the action: stop to stop the environment automatically.

environment:auto_stop_in

Introduced in GitLab 12.8.

The auto_stop_in keyword is for specifying the lifetime of the environment, that when expired, GitLab automatically stops them.

For example,

review_app:
  script: deploy-review-app
  environment:
    name: review/$CI_COMMIT_REF_NAME
    auto_stop_in: 1 day

When the environment for review_app is created, the environment’s lifetime is set to 1 day. Every time the review app is deployed, that lifetime is also reset to 1 day.

For more information, see the environments auto-stop documentation

environment:kubernetes

Introduced in GitLab 12.6.

Use the kubernetes keyword to configure deployments to a Kubernetes cluster that is associated with your project.

For example:

deploy:
  stage: deploy
  script: make deploy-app
  environment:
    name: production
    kubernetes:
      namespace: production

This configuration sets up the deploy job to deploy to the production environment, using the production Kubernetes namespace.

For more information, see Available settings for kubernetes.

NOTE: Kubernetes configuration is not supported for Kubernetes clusters that are managed by GitLab. To follow progress on support for GitLab-managed clusters, see the relevant issue.

environment:deployment_tier

Introduced in GitLab 13.10.

Use the deployment_tier keyword to specify the tier of the deployment environment:

deploy:
  script: echo
  environment:
    name: customer-portal
    deployment_tier: production

For more information, see Deployment tier of environments.

Dynamic environments

Use CI/CD variables to dynamically name environments.

For example:

deploy as review app:
  stage: deploy
  script: make deploy
  environment:
    name: review/$CI_COMMIT_REF_NAME
    url: https://$CI_ENVIRONMENT_SLUG.example.com/

The deploy as review app job is marked as a deployment to dynamically create the review/CI_COMMIT_REF_NAME environment. CI_COMMIT_REF_NAME is a CI/CD variable set by the runner. The

The common use case is to create dynamic environments for branches and use them as Review Apps. You can see an example that uses Review Apps at https://gitlab.com/gitlab-examples/review-apps-nginx/.

cache

Use cache to specify a list of files and directories to cache between jobs. You can only use paths that are in the local working copy.

If cache is defined outside the scope of jobs, it’s set globally and all jobs use that configuration.

Caching is shared between pipelines and jobs. Caches are restored before artifacts.

Read how caching works and find out some good practices in the caching dependencies documentation.

cache:paths

Use the paths directive to choose which files or directories to cache. Paths are relative to the project directory ($CI_PROJECT_DIR) and can’t directly link outside it. You can use Wildcards that use glob patterns and:

• In GitLab Runner 13.0 and later, doublestar.Glob.
• In GitLab Runner 12.10 and earlier, filepath.Match.

Cache all files in binaries that end in .apk and the .config file:

rspec:
  script: test
  cache:
    paths:
      - binaries/*.apk
      - .config

Locally defined cache overrides globally defined options. The following rspec job caches only binaries/:

cache:
  paths:
    - my/files

rspec:
  script: test
  cache:
    key: rspec
    paths:
      - binaries/

The cache is shared between jobs, so if you’re using different paths for different jobs, you should also set a different cache:key. Otherwise cache content can be overwritten.

cache:key

The key keyword defines the affinity of caching between jobs. You can have a single cache for all jobs, cache per-job, cache per-branch, or any other way that fits your workflow. You can fine tune caching, including caching data between different jobs or even different branches.

The cache:key variable can use any of the predefined variables. The default key, if not set, is just literal default, which means everything is shared between pipelines and jobs by default.

For example, to enable per-branch caching:

cache:
  key: "$CI_COMMIT_REF_SLUG"
  paths:
    - binaries/

If you use Windows Batch to run your shell scripts you need to replace $ with %:

cache:
  key: "%CI_COMMIT_REF_SLUG%"
  paths:
    - binaries/

The cache:key variable can’t contain the / character, or the equivalent URI-encoded %2F. A value made only of dots (., %2E) is also forbidden.

You can specify a fallback cache key to use if the specified cache:key is not found.

Multiple caches

• Introduced in GitLab 13.10.
• It’s deployed behind a feature flag, disabled by default.
• It’s disabled on GitLab.com.
• It’s not recommended for production use.
• To use it in GitLab self-managed instances, ask a GitLab administrator to enable it. (FREE SELF)

WARNING: This feature might not be available to you. Check the version history note above for details.

You can have a maximum of four caches:

test-job:
  stage: build
  cache:
    - key:
        files:
          - Gemfile.lock
      paths:
        - vendor/ruby
    - key:
        files:
          - yarn.lock
      paths:
        - .yarn-cache/
  script:
    - bundle install --path=vendor
    - yarn install --cache-folder .yarn-cache
    - echo Run tests...

If multiple caches are combined with a Fallback cache key, the fallback is fetched multiple times if multiple caches are not found.

Enable or disable multiple caches (FREE SELF)

The multiple caches feature is under development and not ready for production use. It is deployed behind a feature flag that is disabled by default. GitLab administrators with access to the GitLab Rails console can enable it.

To enable it:

Feature.enable(:multiple_cache_per_job)

To disable it:

Feature.disable(:multiple_cache_per_job)

Fallback cache key

Introduced in GitLab Runner 13.4.

You can use the CI_COMMIT_REF_SLUG variable to specify your cache:key. For example, if your CI_COMMIT_REF_SLUG is test you can set a job to download cache that’s tagged with test.

If a cache with this tag is not found, you can use CACHE_FALLBACK_KEY to specify a cache to use when none exists.

In the following example, if the $CI_COMMIT_REF_SLUG is not found, the job uses the key defined by the CACHE_FALLBACK_KEY variable:

variables:
  CACHE_FALLBACK_KEY: fallback-key

cache:
  key: "$CI_COMMIT_REF_SLUG"
  paths:
    - binaries/

cache:key:files

Introduced in GitLab v12.5.

The cache:key:files keyword extends the cache:key functionality by making it easier to reuse some caches, and rebuild them less often, which speeds up subsequent pipeline runs.

When you include cache:key:files, you must also list the project files that are used to generate the key, up to a maximum of two files. The cache key is a SHA checksum computed from the most recent commits (up to two, if two files are listed) that changed the given files. If neither file is changed in any commits, the fallback key is default.

cache:
  key:
    files:
      - Gemfile.lock
      - package.json
  paths:
    - vendor/ruby
    - node_modules

This example creates a cache for Ruby and Node.js dependencies that is tied to current versions of the Gemfile.lock and package.json files. Whenever one of these files changes, a new cache key is computed and a new cache is created. Any future job runs that use the same Gemfile.lock and package.json with cache:key:files use the new cache, instead of rebuilding the dependencies.

cache:key:prefix

Introduced in GitLab v12.5.

When you want to combine a prefix with the SHA computed for cache:key:files, use the prefix keyword with key:files. For example, if you add a prefix of test, the resulting key is: test-feef9576d21ee9b6a32e30c5c79d0a0ceb68d1e5. If neither file is changed in any commits, the prefix is added to default, so the key in the example would be test-default.

Like cache:key, prefix can use any of the predefined variables, but cannot include:

• the / character (or the equivalent URI-encoded %2F)
• a value made only of . (or the equivalent URI-encoded %2E)

cache:
  key:
    files:
      - Gemfile.lock
    prefix: ${CI_JOB_NAME}
  paths:
    - vendor/ruby

rspec:
  script:
    - bundle exec rspec

For example, adding a prefix of $CI_JOB_NAME causes the key to look like: rspec-feef9576d21ee9b6a32e30c5c79d0a0ceb68d1e5 and the job cache is shared across different branches. If a branch changes Gemfile.lock, that branch has a new SHA checksum for cache:key:files. A new cache key is generated, and a new cache is created for that key. If Gemfile.lock is not found, the prefix is added to default, so the key in the example would be rspec-default.

cache:untracked

Set untracked: true to cache all files that are untracked in your Git repository:

rspec:
  script: test
  cache:
    untracked: true

Cache all Git untracked files and files in binaries:

rspec:
  script: test
  cache:
    untracked: true
    paths:
      - binaries/

cache:when

Introduced in GitLab 13.5 and GitLab Runner v13.5.0.

cache:when defines when to save the cache, based on the status of the job. You can set cache:when to:

• on_success (default): Save the cache only when the job succeeds.
• on_failure: Save the cache only when the job fails.
• always: Always save the cache.

For example, to store a cache whether or not the job fails or succeeds:

rspec:
  script: rspec
  cache:
    paths:
      - rspec/
    when: 'always'

cache:policy

The default behavior of a caching job is to download the files at the start of execution, and to re-upload them at the end. Any changes made by the job are persisted for future runs. This behavior is known as the pull-push cache policy.

If you know the job does not alter the cached files, you can skip the upload step by setting policy: pull in the job specification. You can add an ordinary cache job at an earlier stage to ensure the cache is updated from time to time:

stages:
  - setup
  - test

prepare:
  stage: setup
  cache:
    key: gems
    paths:
      - vendor/bundle
  script:
    - bundle install --deployment

rspec:
  stage: test
  cache:
    key: gems
    paths:
      - vendor/bundle
    policy: pull
  script:
    - bundle exec rspec ...

Use the pull policy when you have many jobs executing in parallel that use caches. This policy speeds up job execution and reduces load on the cache server.

If you have a job that unconditionally recreates the cache without referring to its previous contents, you can skip the download step. To do so, add policy: push to the job.

artifacts

Use artifacts to specify a list of files and directories that are attached to the job when it succeeds, fails, or always.

The artifacts are sent to GitLab after the job finishes. They are available for download in the GitLab UI if the size is not larger than the maximum artifact size.

By default, jobs in later stages automatically download all the artifacts created by jobs in earlier stages. You can control artifact download behavior in jobs with dependencies.

When using the needs keyword, jobs can only download artifacts from the jobs defined in the needs configuration.

Job artifacts are only collected for successful jobs by default, and artifacts are restored after caches.

Read more about artifacts.

artifacts:paths

Paths are relative to the project directory ($CI_PROJECT_DIR) and can’t directly link outside it. You can use Wildcards that use glob patterns and:

• In GitLab Runner 13.0 and later, doublestar.Glob.
• In GitLab Runner 12.10 and earlier, filepath.Match.

To restrict which jobs a specific job fetches artifacts from, see dependencies.

Send all files in binaries and .config:

artifacts:
  paths:
    - binaries/
    - .config

To disable artifact passing, define the job with empty dependencies:

job:
  stage: build
  script: make build
  dependencies: []

You may want to create artifacts only for tagged releases to avoid filling the build server storage with temporary build artifacts.

Create artifacts only for tags (default-job doesn’t create artifacts):

default-job:
  script:
    - mvn test -U
  except:
    - tags

release-job:
  script:
    - mvn package -U
  artifacts:
    paths:
      - target/*.war
  only:
    - tags

You can use wildcards for directories too. For example, if you want to get all the files inside the directories that end with xyz:

job:
  artifacts:
    paths:
      - path/*xyz/*

artifacts:public

• Introduced in GitLab 13.8
• It’s deployed behind a feature flag, disabled by default.
• It’s enabled on GitLab.com.
• It’s recommended for production use.

Use artifacts:public to determine whether the job artifacts should be publicly available.

The default for artifacts:public is true which means that the artifacts in public pipelines are available for download by anonymous and guest users:

artifacts:
  public: true

To deny read access for anonymous and guest users to artifacts in public pipelines, set artifacts:public to false:

artifacts:
  public: false

artifacts:exclude

• Introduced in GitLab 13.1
• Requires GitLab Runner 13.1

exclude makes it possible to prevent files from being added to an artifacts archive.

Similar to artifacts:paths, exclude paths are relative to the project directory. You can use Wildcards that use glob or filepath.Match patterns.

For example, to store all files in binaries/, but not *.o files located in subdirectories of binaries/:

artifacts:
  paths:
    - binaries/
  exclude:
    - binaries/**/*.o

Files matched by artifacts:untracked can be excluded using artifacts:exclude too.

artifacts:expose_as

Introduced in GitLab 12.5.

Use the expose_as keyword to expose job artifacts in the merge request UI.

For example, to match a single file:

test:
  script: ["echo 'test' > file.txt"]
  artifacts:
    expose_as: 'artifact 1'
    paths: ['file.txt']

With this configuration, GitLab adds a link artifact 1 to the relevant merge request that points to file1.txt. To access the link, select View exposed artifact below the pipeline graph in the merge request overview.

An example that matches an entire directory:

test:
  script: ["mkdir test && echo 'test' > test/file.txt"]
  artifacts:
    expose_as: 'artifact 1'
    paths: ['test/']

Note the following:

• Artifacts do not display in the merge request UI when using variables to define the artifacts:paths.
• A maximum of 10 job artifacts per merge request can be exposed.
• Glob patterns are unsupported.
• If a directory is specified, the link is to the job artifacts browser if there is more than one file in the directory.
• For exposed single file artifacts with .html, .htm, .txt, .json, .xml, and .log extensions, if GitLab Pages is:
  • Enabled, GitLab automatically renders the artifact.
  • Not enabled, the file is displayed in the artifacts browser.

artifacts:name

Use the name directive to define the name of the created artifacts archive. You can specify a unique name for every archive. The artifacts:name variable can make use of any of the predefined variables. The default name is artifacts, which becomes artifacts.zip when you download it.

To create an archive with a name of the current job:

job:
  artifacts:
    name: "$CI_JOB_NAME"
    paths:
      - binaries/

To create an archive with a name of the current branch or tag including only the binaries directory:

job:
  artifacts:
    name: "$CI_COMMIT_REF_NAME"
    paths:
      - binaries/

If your branch-name contains forward slashes (for example feature/my-feature) it’s advised to use CI_COMMIT_REF_SLUG instead of CI_COMMIT_REF_NAME for proper naming of the artifact.

To create an archive with a name of the current job and the current branch or tag including only the binaries directory:

job:
  artifacts:
    name: "$CI_JOB_NAME-$CI_COMMIT_REF_NAME"
    paths:
      - binaries/

To create an archive with a name of the current stage and branch name:

job:
  artifacts:
    name: "$CI_JOB_STAGE-$CI_COMMIT_REF_NAME"
    paths:
      - binaries/

If you use Windows Batch to run your shell scripts you need to replace $ with %:

job:
  artifacts:
    name: "%CI_JOB_STAGE%-%CI_COMMIT_REF_NAME%"
    paths:
      - binaries/

If you use Windows PowerShell to run your shell scripts you need to replace with env::

job:
  artifacts:
    name: "$env:CI_JOB_STAGE-$env:CI_COMMIT_REF_NAME"
    paths:
      - binaries/

artifacts:untracked

Use artifacts:untracked to add all Git untracked files as artifacts (along with the paths defined in artifacts:paths). artifacts:untracked ignores configuration in the repository’s .gitignore file.

Send all Git untracked files:

artifacts:
  untracked: true

Send all Git untracked files and files in binaries:

artifacts:
  untracked: true
  paths:
    - binaries/

Send all untracked files but exclude *.txt:

artifacts:
  untracked: true
  exclude:
    - "*.txt"

artifacts:when

Use artifacts:when to upload artifacts on job failure or despite the failure.

artifacts:when can be set to one of the following values:

1. on_success (default): Upload artifacts only when the job succeeds.
2. on_failure: Upload artifacts only when the job fails.
3. always: Always upload artifacts.

For example, to upload artifacts only when a job fails:

job:
  artifacts:
    when: on_failure

artifacts:expire_in

Use expire_in to specify how long artifacts are active before they expire and are deleted.

The expiration time period begins when the artifact is uploaded and stored on GitLab. If the expiry time is not defined, it defaults to the instance wide setting (30 days by default).

To override the expiration date and protect artifacts from being automatically deleted:

• Use the Keep button on the job page.
• Set the value of expire_in to never. Available in GitLab 13.3 and later.

After their expiry, artifacts are deleted hourly by default (via a cron job), and are not accessible anymore.

The value of expire_in is an elapsed time in seconds, unless a unit is provided. Examples of valid values:

• '42'
• 42 seconds
• 3 mins 4 sec
• 2 hrs 20 min
• 2h20min
• 6 mos 1 day
• 47 yrs 6 mos and 4d
• 3 weeks and 2 days
• never

To expire artifacts 1 week after being uploaded:

job:
  artifacts:
    expire_in: 1 week

The latest artifacts for refs are locked against deletion, and kept regardless of the expiry time. Introduced in GitLab 13.0 behind a disabled feature flag, and made the default behavior in GitLab 13.4.

In GitLab 13.8 and later, you can disable this behavior at the project level in the CI/CD settings. In GitLab 13.9 and later, you can disable this behavior instance-wide.

artifacts:reports

Use artifacts:reports to collect test reports, code quality reports, and security reports from jobs. It also exposes these reports in the GitLab UI (merge requests, pipeline views, and security dashboards).

These are the available report types:

dependencies

By default, all artifacts from previous stages are passed to each job. However, you can use the dependencies keyword to define a limited list of jobs to fetch artifacts from. You can also set a job to download no artifacts at all.

To use this feature, define dependencies in context of the job and pass a list of all previous jobs the artifacts should be downloaded from.

You can define jobs from stages that were executed before the current one. An error occurs if you define jobs from the current or an upcoming stage.

To prevent a job from downloading artifacts, define an empty array.

When you use dependencies, the status of the previous job is not considered. If a job fails or it’s a manual job that isn’t triggered, no error occurs.

The following example defines two jobs with artifacts: build:osx and build:linux. When the test:osx is executed, the artifacts from build:osx are downloaded and extracted in the context of the build. The same happens for test:linux and artifacts from build:linux.

The job deploy downloads artifacts from all previous jobs because of the stage precedence:

build:osx:
  stage: build
  script: make build:osx
  artifacts:
    paths:
      - binaries/

build:linux:
  stage: build
  script: make build:linux
  artifacts:
    paths:
      - binaries/

test:osx:
  stage: test
  script: make test:osx
  dependencies:
    - build:osx

test:linux:
  stage: test
  script: make test:linux
  dependencies:
    - build:linux

deploy:
  stage: deploy
  script: make deploy

When a dependent job fails

Introduced in GitLab 10.3.

If the artifacts of the job that is set as a dependency are expired or erased, then the dependent job fails.

You can ask your administrator to flip this switch and bring back the old behavior.

coverage

Use coverage to configure how code coverage is extracted from the job output.

Regular expressions are the only valid kind of value expected here. So, using surrounding / is mandatory to consistently and explicitly represent a regular expression string. You must escape special characters if you want to match them literally.

For example:

job1:
  script: rspec
  coverage: '/Code coverage: \d+\.\d+/'

The coverage is shown in the UI if at least one line in the job output matches the regular expression. If there is more than one matched line in the job output, the last line is used. For the matched line, the first occurrence of \d+(\.\d+)? is the code coverage. Leading zeros are removed.

Coverage output from child pipelines is not recorded or displayed. Check the related issue for more details.

retry

Introduced in GitLab 11.5, you can control which failures to retry on.

Use retry to configure how many times a job is retried in case of a failure.

When a job fails, the job is processed again, until the limit specified by the retry keyword is reached.

If retry is set to 2, and a job succeeds in a second run (first retry), it is not retried. The retry value must be a positive integer, from 0 to 2 (two retries maximum, three runs in total).

The following example retries all failure cases:

test:
  script: rspec
  retry: 2

By default, a job is retried on all failure cases. To have better control over which failures to retry, retry can be a hash with the following keys:

• max: The maximum number of retries.
• when: The failure cases to retry.

To retry only runner system failures at maximum two times:

test:
  script: rspec
  retry:
    max: 2
    when: runner_system_failure

If there is another failure, other than a runner system failure, the job is not retried.

To retry on multiple failure cases, when can also be an array of failures:

test:
  script: rspec
  retry:
    max: 2
    when:
      - runner_system_failure
      - stuck_or_timeout_failure

Possible values for when are:

• always: Retry on any failure (default).
• unknown_failure: Retry when the failure reason is unknown.
• script_failure: Retry when the script failed.
• api_failure: Retry on API failure.
• stuck_or_timeout_failure: Retry when the job got stuck or timed out.
• runner_system_failure: Retry if there is a runner system failure (for example, job setup failed).
• missing_dependency_failure: Retry if a dependency is missing.
• runner_unsupported: Retry if the runner is unsupported.
• stale_schedule: Retry if a delayed job could not be executed.
• job_execution_timeout: Retry if the script exceeded the maximum execution time set for the job.
• archived_failure: Retry if the job is archived and can’t be run.
• unmet_prerequisites: Retry if the job failed to complete prerequisite tasks.
• scheduler_failure: Retry if the scheduler failed to assign the job to a runner.
• data_integrity_failure: Retry if there is a structural integrity problem detected.

You can specify the number of retry attempts for certain stages of job execution using variables.

timeout

Introduced in GitLab 12.3.

Use timeout to configure a timeout for a specific job. For example:

build:
  script: build.sh
  timeout: 3 hours 30 minutes

test:
  script: rspec
  timeout: 3h 30m

The job-level timeout can exceed the project-level timeout but can’t exceed the runner-specific timeout.

parallel

Introduced in GitLab 11.5.

Use parallel to configure how many instances of a job to run in parallel. The value can be from 2 to 50.

The parallel keyword creates N instances of the same job that run in parallel. They are named sequentially from job_name 1/N to job_name N/N:

test:
  script: rspec
  parallel: 5

Every parallel job has a CI_NODE_INDEX and CI_NODE_TOTAL predefined CI/CD variable set.

Different languages and test suites have different methods to enable parallelization. For example, use Semaphore Test Boosters and RSpec to run Ruby tests in parallel:

# Gemfile
source 'https://rubygems.org'

gem 'rspec'
gem 'semaphore_test_boosters'

test:
  parallel: 3
  script:
    - bundle
    - bundle exec rspec_booster --job $CI_NODE_INDEX/$CI_NODE_TOTAL

WARNING: Test Boosters reports usage statistics to the author.

You can then navigate to the Jobs tab of a new pipeline build and see your RSpec job split into three separate jobs.

Parallel matrix jobs

• Introduced in GitLab 13.3.

Use matrix: to run a job multiple times in parallel in a single pipeline, but with different variable values for each instance of the job. There can be from 2 to 50 jobs.

Jobs can only run in parallel if there are multiple runners, or a single runner is configured to run multiple jobs concurrently.

Every job gets the same CI_NODE_TOTAL CI/CD variable value, and a unique CI_NODE_INDEX value.

deploystacks:
  stage: deploy
  script:
    - bin/deploy
  parallel:
    matrix:
      - PROVIDER: aws
        STACK:
          - monitoring
          - app1
          - app2
      - PROVIDER: ovh
        STACK: [monitoring, backup, app]
      - PROVIDER: [gcp, vultr]
        STACK: [data, processing]

The following example generates 10 parallel deploystacks jobs, each with different values for PROVIDER and STACK:

deploystacks: [aws, monitoring]
deploystacks: [aws, app1]
deploystacks: [aws, app2]
deploystacks: [ovh, monitoring]
deploystacks: [ovh, backup]
deploystacks: [ovh, app]
deploystacks: [gcp, data]
deploystacks: [gcp, processing]
deploystacks: [vultr, data]
deploystacks: [vultr, processing]

The job naming style was improved in GitLab 13.4.

One-dimensional matrix jobs

Introduced in GitLab 13.5.

You can also have one-dimensional matrices with a single job:

deploystacks:
  stage: deploy
  script:
    - bin/deploy
  parallel:
    matrix:
      - PROVIDER: [aws, ovh, gcp, vultr]

Parallel matrix trigger jobs

Introduced in GitLab 13.10.

Use matrix: to run a trigger job multiple times in parallel in a single pipeline, but with different variable values for each instance of the job.

deploystacks:
  stage: deploy
  trigger:
    include: path/to/child-pipeline.yml
  parallel:
    matrix:
      - PROVIDER: aws
        STACK: [monitoring, app1]
      - PROVIDER: ovh
        STACK: [monitoring, backup]
      - PROVIDER: [gcp, vultr]
        STACK: [data]

This example generates 6 parallel deploystacks trigger jobs, each with different values for PROVIDER and STACK, and they create 6 different child pipelines with those variables.

deploystacks: [aws, monitoring]
deploystacks: [aws, app1]
deploystacks: [ovh, monitoring]
deploystacks: [ovh, backup]
deploystacks: [gcp, data]
deploystacks: [vultr, data]

trigger

• Introduced in GitLab Premium 11.8.
• Moved to GitLab Free in 12.8.

Use trigger to define a downstream pipeline trigger. When GitLab starts a trigger job, a downstream pipeline is created.

Jobs with trigger can only use a limited set of keywords. For example, you can’t run commands with script, before_script, or after_script.

You can use this keyword to create two different types of downstream pipelines:

• Multi-project pipelines
• Child pipelines

In GitLab 13.2 and later, you can view which job triggered a downstream pipeline. In the pipeline graph, hover over the downstream pipeline job.

In GitLab 13.5 and later, you can use when:manual in the same job as trigger. In GitLab 13.4 and earlier, using them together causes the error jobs:#{job-name} when should be on_success, on_failure or always. You cannot start manual trigger jobs with the API.

Basic trigger syntax for multi-project pipelines

You can configure a downstream trigger by using the trigger keyword with a full path to a downstream project:

rspec:
  stage: test
  script: bundle exec rspec

staging:
  stage: deploy
  trigger: my/deployment

Complex trigger syntax for multi-project pipelines

You can configure a branch name that GitLab uses to create a downstream pipeline with:

rspec:
  stage: test
  script: bundle exec rspec

staging:
  stage: deploy
  trigger:
    project: my/deployment
    branch: stable

To mirror the status from a triggered pipeline:

trigger_job:
  trigger:
    project: my/project
    strategy: depend

To mirror the status from an upstream pipeline:

upstream_bridge:
  stage: test
  needs:
    pipeline: other/project

trigger syntax for child pipeline

Introduced in GitLab 12.7.

To create a child pipeline, specify the path to the YAML file that contains the configuration of the child pipeline:

trigger_job:
  trigger:
    include: path/to/child-pipeline.yml

Similar to multi-project pipelines, it’s possible to mirror the status from a triggered pipeline:

trigger_job:
  trigger:
    include:
      - local: path/to/child-pipeline.yml
    strategy: depend

Trigger child pipeline with generated configuration file

Introduced in GitLab 12.9.

You can also trigger a child pipeline from a dynamically generated configuration file:

generate-config:
  stage: build
  script: generate-ci-config > generated-config.yml
  artifacts:
    paths:
      - generated-config.yml

child-pipeline:
  stage: test
  trigger:
    include:
      - artifact: generated-config.yml
        job: generate-config

The generated-config.yml is extracted from the artifacts and used as the configuration for triggering the child pipeline.

Trigger child pipeline with files from another project

Introduced in GitLab 13.5.

To trigger child pipelines with files from another private project under the same GitLab instance, use include:file:

child-pipeline:
  trigger:
    include:
      - project: 'my-group/my-pipeline-library'
        ref: 'master'
        file: '/path/to/child-pipeline.yml'

Linking pipelines with trigger:strategy

By default, the trigger job completes with the success status as soon as the downstream pipeline is created.

To force the trigger job to wait for the downstream (multi-project or child) pipeline to complete, use strategy: depend. This setting makes the trigger job wait with a “running” status until the triggered pipeline completes. At that point, the trigger job completes and displays the same status as the downstream job.

This setting can help keep your pipeline execution linear. In the following example, jobs from subsequent stages wait for the triggered pipeline to successfully complete before starting, which reduces parallelization.

trigger_job:
  trigger:
    include: path/to/child-pipeline.yml
    strategy: depend

Trigger a pipeline by API call

To force a rebuild of a specific branch, tag, or commit, you can use an API call with a trigger token.

The trigger token is different than the trigger keyword.

Read more in the triggers documentation.

interruptible

Introduced in GitLab 12.3.

Use interruptible to indicate that a running job should be canceled if made redundant by a newer pipeline run. Defaults to false (uninterruptible). Jobs that have not started yet (pending) are considered interruptible and safe to be cancelled. This value is used only if the automatic cancellation of redundant pipelines feature is enabled.

When enabled, a pipeline is immediately canceled when a new pipeline starts on the same branch if either of the following is true:

• All jobs in the pipeline are set as interruptible.
• Any uninterruptible jobs have not started yet.

Set jobs as interruptible that can be safely canceled once started (for instance, a build job).

In the following example, a new pipeline run causes an existing running pipeline to be:

• Canceled, if only step-1 is running or pending.
• Not canceled, once step-2 starts running.

After an uninterruptible job starts running, the pipeline cannot be canceled.

stages:
  - stage1
  - stage2
  - stage3

step-1:
  stage: stage1
  script:
    - echo "Can be canceled."
  interruptible: true

step-2:
  stage: stage2
  script:
    - echo "Can not be canceled."

step-3:
  stage: stage3
  script:
    - echo "Because step-2 can not be canceled, this step can never be canceled, even though it's set as interruptible."
  interruptible: true

resource_group

Introduced in GitLab 12.7.

Sometimes running multiple jobs or pipelines at the same time in an environment can lead to errors during the deployment.

To avoid these errors, use the resource_group attribute to make sure that the runner doesn’t run certain jobs simultaneously. Resource groups behave similar to semaphores in other programming languages.

When the resource_group keyword is defined for a job in the .gitlab-ci.yml file, job executions are mutually exclusive across different pipelines for the same project. If multiple jobs belonging to the same resource group are enqueued simultaneously, only one of the jobs is picked by the runner. The other jobs wait until the resource_group is free.

For example:

deploy-to-production:
  script: deploy
  resource_group: production

In this case, two deploy-to-production jobs in two separate pipelines can never run at the same time. As a result, you can ensure that concurrent deployments never happen to the production environment.

You can define multiple resource groups per environment. For example, when deploying to physical devices, you may have multiple physical devices. Each device can be deployed to, but there can be only one deployment per device at any given time.

The resource_group value can only contain letters, digits, -, _, /, $, {, }, ., and spaces. It can’t start or end with /.

For more information, see Deployments Safety.

Pipeline-level concurrency control with Cross-Project/Parent-Child pipelines

Introduced in GitLab 13.9.

You can define resource_group for downstream pipelines that are sensitive to concurrent executions. The trigger keyword can trigger downstream pipelines. The resource_group keyword can co-exist with it. This is useful to control the concurrency for deployment pipelines, while running non-sensitive jobs concurrently.

The following example has two pipeline configurations in a project. When a pipeline starts running, non-sensitive jobs are executed first and aren’t affected by concurrent executions in other pipelines. However, GitLab ensures that there are no other deployment pipelines running before triggering a deployment (child) pipeline. If other deployment pipelines are running, GitLab waits until those pipelines finish before running another one.

# .gitlab-ci.yml (parent pipeline)

build:
  stage: build
  script: echo "Building..."

test:
  stage: test
  script: echo "Testing..."

deploy:
  stage: deploy
  trigger:
    include: deploy.gitlab-ci.yml
    strategy: depend
  resource_group: AWS-production

# deploy.gitlab-ci.yml (child pipeline)

stages:
  - provision
  - deploy

provision:
  stage: provision
  script: echo "Provisioning..."

deployment:
  stage: deploy
  script: echo "Deploying..."

You must define strategy: depend with the trigger keyword. This ensures that the lock isn’t released until the downstream pipeline finishes.

release

Introduced in GitLab 13.2.

Use release to create a release. Requires the release-cli to be available in your GitLab Runner Docker or shell executor.

These keywords are supported:

• tag_name
• description
• name (optional)
• ref (optional)
• milestones (optional)
• released_at (optional)

The release is created only if the job processes without error. If the Rails API returns an error during release creation, the release job fails.

release-cli Docker image

You must specify the Docker image to use for the release-cli:

image: registry.gitlab.com/gitlab-org/release-cli:latest

release-cli for shell executors

Introduced in GitLab 13.8.

For GitLab Runner shell executors, you can download and install the release-cli manually for your supported OS and architecture. Once installed, the release keyword should be available to you.

Install on Unix/Linux

1. Download the binary for your system, in the following example for amd64 systems:

curl --location --output /usr/local/bin/release-cli "https://release-cli-downloads.s3.amazonaws.com/latest/release-cli-linux-amd64"

1. Give it permissions to execute:

sudo chmod +x /usr/local/bin/release-cli

1. Verify release-cli is available:

$ release-cli -v

release-cli version 0.6.0

Install on Windows PowerShell

1. Create a folder somewhere in your system, for example C:\GitLab\Release-CLI\bin

New-Item -Path 'C:\GitLab\Release-CLI\bin' -ItemType Directory

1. Download the executable file:

PS C:\> Invoke-WebRequest -Uri "https://release-cli-downloads.s3.amazonaws.com/latest/release-cli-windows-amd64.exe" -OutFile "C:\GitLab\Release-CLI\bin\release-cli.exe"

    Directory: C:\GitLab\Release-CLI
Mode                LastWriteTime         Length Name
----                -------------         ------ ----
d-----        3/16/2021   4:17 AM                bin

1. Add the directory to your $env:PATH:

$env:PATH += ";C:\GitLab\Release-CLI\bin"

1. Verify release-cli is available:

PS C:\> release-cli -v

release-cli version 0.6.0

script

All jobs except trigger jobs must have the script keyword. A release job can use the output from script commands, but you can use a placeholder script if the script is not needed:

script:
  - echo 'release job'

An issue exists to remove this requirement in an upcoming version of GitLab.

A pipeline can have multiple release jobs, for example:

ios-release:
  script:
    - echo 'iOS release job'
  release:
    tag_name: v1.0.0-ios
    description: 'iOS release v1.0.0'

android-release:
  script:
    - echo 'Android release job'
  release:
    tag_name: v1.0.0-android
    description: 'Android release v1.0.0'

release:tag_name

You must specify a tag_name for the release. The tag can refer to an existing Git tag or you can specify a new tag.

When the specified tag doesn’t exist in the repository, a new tag is created from the associated SHA of the pipeline.

For example, when creating a release from a Git tag:

job:
  release:
    tag_name: $CI_COMMIT_TAG
    description: 'Release description'

It is also possible to create any unique tag, in which case only: tags is not mandatory. A semantic versioning example:

job:
  release:
    tag_name: ${MAJOR}_${MINOR}_${REVISION}
    description: 'Release description'

• The release is created only if the job’s main script succeeds.
• If the release already exists, it is not updated and the job with the release keyword fails.
• The release section executes after the script tag and before the after_script.

release:name

The release name. If omitted, it is populated with the value of release: tag_name.

release:description

Specifies the long description of the release. You can also specify a file that contains the description.

Read description from a file

Introduced in GitLab 13.7.

You can specify a file in CI_PROJECT_DIR that contains the description. The file must be relative to the project directory (CI_PROJECT_DIR), and if the file is a symbolic link it can’t reside outside of

job:
  release:
    tag_name: ${MAJOR}_${MINOR}_${REVISION}
    description: './path/to/CHANGELOG.md'

release:ref

If the release: tag_name doesn’t exist yet, the release is created from ref. ref can be a commit SHA, another tag name, or a branch name.

release:milestones

The title of each milestone the release is associated with.

release:released_at

The date and time when the release is ready. Defaults to the current date and time if not defined. Should be enclosed in quotes and expressed in ISO 8601 format.

released_at: '2021-03-15T08:00:00Z'

Complete example for release

If you combine the previous examples for release, you get two options, depending on how you generate the tags. You can’t use these options together, so choose one:

Release assets as Generic packages

You can use Generic packages to host your release assets. For a complete example, see the Release assets as Generic packages project.

release-cli command line

The entries under the release node are transformed into a bash command line and sent to the Docker container, which contains the release-cli. You can also call the release-cli directly from a script entry.

For example, if you use the YAML described previously:

release-cli create --name "Release $CI_COMMIT_SHA" --description "Created using the release-cli $EXTRA_DESCRIPTION" --tag-name "v${MAJOR}.${MINOR}.${REVISION}" --ref "$CI_COMMIT_SHA" --released-at "2020-07-15T08:00:00Z" --milestone "m1" --milestone "m2" --milestone "m3"

secrets

Introduced in GitLab 13.4.

Use secrets to specify the CI/CD Secrets the job needs. It should be a hash, and the keys should be the names of the variables that are made available to the job. The value of each secret is saved in a temporary file. This file’s path is stored in these variables.

secrets:vault (PREMIUM)

Introduced in GitLab 13.4.

Use vault to specify secrets provided by Hashicorp’s Vault.

This syntax has multiple forms. The shortest form assumes the use of the KV-V2 secrets engine, mounted at the default path kv-v2. The last part of the secret’s path is the field to fetch the value for:

job:
  secrets:
    DATABASE_PASSWORD:
      vault: production/db/password  # translates to secret `kv-v2/data/production/db`, field `password`

You can specify a custom secrets engine path by adding a suffix starting with @:

job:
  secrets:
    DATABASE_PASSWORD:
      vault: production/db/password@ops  # translates to secret `ops/data/production/db`, field `password`

In the detailed form of the syntax, you can specify all details explicitly:

job:
  secrets:
    DATABASE_PASSWORD:      # translates to secret `ops/data/production/db`, field `password`
      vault:
        engine:
          name: kv-v2
          path: ops
        path: production/db
        field: password

pages

Use pages to upload static content to GitLab. The content is then published as a website. You must:

• Place any static content in a public/ directory.
• Define artifacts with a path to the public/ directory.

The following example moves all files from the root of the project to the public/ directory. The .public workaround is so cp does not also copy public/ to itself in an infinite loop:

pages:
  stage: deploy
  script:
    - mkdir .public
    - cp -r * .public
    - mv .public public
  artifacts:
    paths:
      - public
  only:
    - master

View the GitLab Pages user documentation.

inherit

Introduced in GitLab 12.9

Use inherit: to control inheritance of globally-defined defaults and variables.

To enable or disable the inheritance of all default: or variables: keywords, use:

• default: true or default: false
• variables: true or variables: false

To inherit only a subset of default: keywords or variables:, specify what you wish to inherit. Anything not listed is not inherited. Use one of the following formats:

inherit:
  default: [keyword1, keyword2]
  variables: [VARIABLE1, VARIABLE2]

Or:

inherit:
  default:
    - keyword1
    - keyword2
  variables:
    - VARIABLE1
    - VARIABLE2

In the following example:

• rubocop:
  • inherits: Nothing.
• rspec:
  • inherits: the default image and the WEBHOOK_URL variable.
  • does not inherit: the default before_script and the DOMAIN variable.
• capybara:
  • inherits: the default before_script and image.
  • does not inherit: the DOMAIN and WEBHOOK_URL variables.
• karma:
  • inherits: the default image and before_script, and the DOMAIN variable.
  • does not inherit: WEBHOOK_URL variable.

default:
  image: 'ruby:2.4'
  before_script:
    - echo Hello World

variables:
  DOMAIN: example.com
  WEBHOOK_URL: https://my-webhook.example.com

rubocop:
  inherit:
    default: false
    variables: false
  script: bundle exec rubocop

rspec:
  inherit:
    default: [image]
    variables: [WEBHOOK_URL]
  script: bundle exec rspec

capybara:
  inherit:
    variables: false
  script: bundle exec capybara

karma:
  inherit:
    default: true
    variables: [DOMAIN]
  script: karma

variables

Introduced in GitLab Runner v0.5.0.

CI/CD variables are configurable values that are passed to jobs. They can be set globally and per-job.

There are two types of variables.

• Custom variables: You can define their values in the .gitlab-ci.yml file, in the GitLab UI, or by using the API. You can also input variables in the GitLab UI when running a pipeline manually.
• Predefined variables: These values are set by the runner itself. One example is CI_COMMIT_REF_NAME, which is the branch or tag the project is built for.

After you define a variable, you can use it in all executed commands and scripts.

Variables are meant for non-sensitive project configuration, for example:

variables:
  DEPLOY_SITE: "https://example.com/"

deploy_job:
  stage: deploy
  script:
    - deploy-script --url $DEPLOY_SITE --path "/"

deploy_review_job:
  stage: deploy
  variables:
    REVIEW_PATH: "/review"
  script:
    - deploy-review-script --url $DEPLOY_SITE --path $REVIEW_PATH

You can use only integers and strings for the variable’s name and value.

If you define a variable at the top level of the gitlab-ci.yml file, it is global, meaning it applies to all jobs. If you define a variable in a job, it’s available to that job only.

If a variable of the same name is defined globally and for a specific job, the job-specific variable overrides the global variable.

All YAML-defined variables are also set to any linked Docker service containers.

You can use YAML anchors for variables.

Prefill variables in manual pipelines

Introduced in GitLab 13.7.

Use the value and description keywords to define variables that are prefilled when running a pipeline manually:

variables:
  DEPLOY_ENVIRONMENT:
    value: "staging"  # Deploy to staging by default
    description: "The deployment target. Change this variable to 'canary' or 'production' if needed."

Configure runner behavior with variables

You can use CI/CD variables to configure how the runner processes Git requests:

• GIT_STRATEGY
• GIT_SUBMODULE_STRATEGY
• GIT_CHECKOUT
• GIT_CLEAN_FLAGS
• GIT_FETCH_EXTRA_FLAGS
• GIT_DEPTH (shallow cloning)
• GIT_CLONE_PATH (custom build directories)
• TRANSFER_METER_FREQUENCY (artifact/cache meter update frequency)
• ARTIFACT_COMPRESSION_LEVEL (artifact archiver compression level)
• CACHE_COMPRESSION_LEVEL (cache archiver compression level)

You can also use variables to configure how many times a runner attempts certain stages of job execution.

YAML-specific features

In your .gitlab-ci.yml file, you can use YAML-specific features like anchors (&), aliases (*), and map merging (<<). Use these features to reduce the complexity of the code in the .gitlab-ci.yml file.

Read more about the various YAML features.

In most cases, the extends keyword is more user friendly and you should use it when possible.

You can use YAML anchors to merge YAML arrays.

Anchors

YAML has a feature called ‘anchors’ that you can use to duplicate content across your document.

Use anchors to duplicate or inherit properties. Use anchors with hidden jobs to provide templates for your jobs. When there are duplicate keys, GitLab performs a reverse deep merge based on the keys.

You can’t use YAML anchors across multiple files when using the include keyword. Anchors are only valid in the file they were defined in. To reuse configuration from different YAML files, use !reference tags or the extends keyword.

The following example uses anchors and map merging. It creates two jobs, test1 and test2, that inherit the .job_template configuration, each with their own custom script defined:

.job_template: &job_configuration  # Hidden yaml configuration that defines an anchor named 'job_configuration'
  image: ruby:2.6
  services:
    - postgres
    - redis

test1:
  <<: *job_configuration           # Merge the contents of the 'job_configuration' alias
  script:
    - test1 project

test2:
  <<: *job_configuration           # Merge the contents of the 'job_configuration' alias
  script:
    - test2 project

& sets up the name of the anchor (job_configuration), << means “merge the given hash into the current one,” and * includes the named anchor (job_configuration again). The expanded version of this example is:

.job_template:
  image: ruby:2.6
  services:
    - postgres
    - redis

test1:
  image: ruby:2.6
  services:
    - postgres
    - redis
  script:
    - test1 project

test2:
  image: ruby:2.6
  services:
    - postgres
    - redis
  script:
    - test2 project

You can use anchors to define two sets of services. For example, test:postgres and test:mysql share the script defined in .job_template, but use different services, defined in .postgres_services and .mysql_services:

.job_template: &job_configuration
  script:
    - test project
  tags:
    - dev

.postgres_services:
  services: &postgres_configuration
    - postgres
    - ruby

.mysql_services:
  services: &mysql_configuration
    - mysql
    - ruby

test:postgres:
  <<: *job_configuration
  services: *postgres_configuration
  tags:
    - postgres

test:mysql:
  <<: *job_configuration
  services: *mysql_configuration

The expanded version is:

.job_template:
  script:
    - test project
  tags:
    - dev

.postgres_services:
  services:
    - postgres
    - ruby

.mysql_services:
  services:
    - mysql
    - ruby

test:postgres:
  script:
    - test project
  services:
    - postgres
    - ruby
  tags:
    - postgres

test:mysql:
  script:
    - test project
  services:
    - mysql
    - ruby
  tags:
    - dev

You can see that the hidden jobs are conveniently used as templates, and tags: [postgres] overwrites tags: [dev].

YAML anchors for scripts

Introduced in GitLab 12.5.

You can use YAML anchors with script, before_script, and after_script to use predefined commands in multiple jobs:

.some-script-before: &some-script-before
  - echo "Execute this script first"

.some-script: &some-script
  - echo "Execute this script second"
  - echo "Execute this script too"

.some-script-after: &some-script-after
  - echo "Execute this script last"

job1:
  before_script:
    - *some-script-before
  script:
    - *some-script
    - echo "Execute something, for this job only"
  after_script:
    - *some-script-after

job2:
  script:
    - *some-script-before
    - *some-script
    - echo "Execute something else, for this job only"
    - *some-script-after

YAML anchors for variables

Use YAML anchors with variables to repeat assignment of variables across multiple jobs. You can also use YAML anchors when a job requires a specific variables block that would otherwise override the global variables.

The following example shows how override the GIT_STRATEGY variable without affecting the use of the SAMPLE_VARIABLE variable:

# global variables
variables: &global-variables
  SAMPLE_VARIABLE: sample_variable_value
  ANOTHER_SAMPLE_VARIABLE: another_sample_variable_value

# a job that must set the GIT_STRATEGY variable, yet depend on global variables
job_no_git_strategy:
  stage: cleanup
  variables:
    <<: *global-variables
    GIT_STRATEGY: none
  script: echo $SAMPLE_VARIABLE

Hide jobs

If you want to temporarily disable a job, rather than commenting out all the lines where the job is defined:

# hidden_job:
#   script:
#     - run test

Instead, you can start its name with a dot (.) and it is not processed by GitLab CI/CD. In the following example, .hidden_job is ignored:

.hidden_job:
  script:
    - run test

Use this feature to ignore jobs, or use the YAML-specific features and transform the hidden jobs into templates.

!reference tags

• Introduced in GitLab 13.9.

Use the !reference custom YAML tag to select keyword configuration from other job sections and reuse it in the current section. Unlike YAML anchors, you can use !reference tags to reuse configuration from included configuration files as well.

In the following example, a script and an after_script from two different locations are reused in the test job:

In the following example, test-vars-1 reuses the all the variables in .vars, while test-vars-2 selects a specific variable and reuses it as a new MY_VAR variable.

.vars:
  variables:
    URL: "http://my-url.internal"
    IMPORTANT_VAR: "the details"

test-vars-1:
  variables: !reference [.vars, variables]
  script:
    - printenv

test-vars-2:
  variables:
    MY_VAR: !reference [.vars, variables, IMPORTANT_VAR]
  script:
    - printenv

You can’t reuse a section that already includes a !reference tag. Only one level of nesting is supported.

Skip Pipeline

To push a commit without triggering a pipeline, add [ci skip] or [skip ci], using any capitalization, to your commit message.

Alternatively, if you are using Git 2.10 or later, use the ci.skip Git push option. The ci.skip push option does not skip merge request pipelines.

Processing Git pushes

GitLab creates at most four branch and tag pipelines when pushing multiple changes in a single git push invocation.

This limitation does not affect any of the updated merge request pipelines. All updated merge requests have a pipeline created when using pipelines for merge requests.

Deprecated keywords

The following keywords are deprecated.

Globally-defined types

WARNING: types is deprecated, and could be removed in a future release. Use stages instead.

Job-defined type

WARNING: type is deprecated, and could be removed in one of the future releases. Use stage instead.

Globally-defined image, services, cache, before_script, after_script

Defining image, services, cache, before_script, and after_script globally is deprecated. Support could be removed from a future release.

Use default: instead. For example:

default:
  image: ruby:2.5
  services:
    - docker:dind
  cache:
    paths: [vendor/]
  before_script:
    - bundle install --path vendor/
  after_script:
    - rm -rf tmp/