Skip to content

kachick/wait-other-jobs

Use this GitHub action with your project
Add this Action to an existing workflow or create a new one
View on Marketplace

Repository files navigation

wait-other-jobs

Itself
gracePeriod eventName matrix
TypeScript

This GitHub action waits for all or specific jobs, even if they are running in other workflows.
If any of those jobs fail, this action will fail as well.

Usage

Zero or tiny configuration may work.

jobs:
  your_job:
    # Enabling these permissions are required in private repositories
    # permissions:
    #   contents: read
    #   checks: read
    #   actions: read
    runs-on: ubuntu-24.04
    steps:
      - uses: kachick/[email protected]
        timeout-minutes: 15 # Recommended to be enabled with your appropriate value for fail-safe use

You can change the token, polling interval, allow/deny list and turns early-exit as below.

with:
  github-token: '${{ secrets.YOUR_PAT }}'
  wait-seconds-before-first-polling: '30' # default '10'
  min-interval-seconds: '300' # default '15'
  retry-method: 'exponential_backoff' # default 'equal_intervals'
  early-exit: 'false' # default 'true'
  skip-same-workflow: 'true' # default 'false'
  # lists should be given with JSON formatted array, do not specify both wait-list and skip-list
  #   - Each item should have a "workflowFile" field, and they can optionally have a "jobName" field.
  #   - If no jobName is specified, all the jobs in the workflow will be targeted.
  #   - wait-list:
  #     - If the checkRun for the specified name is not found, this action raise errors by default.
  #       You can disable this validation with `"optional": true` or use the `startupGracePeriod` that described in following section
  #     - Wait for all event types by default, you can change with `"eventName": "EVENT_NAME_AS_push"`.
  wait-list: |
    [
      {
        "workflowFile": "ci.yml",
        "jobName": "test",
        "eventName": "${{ github.event_name }}"
      },
      {
        "workflowFile": "release.yml",
        "optional": true
      }
    ]
  skip-list: |
    [
      {
        "workflowFile": "pages.yml"
      }
    ]

Full list of the options

NAME DESCRIPTION TYPE DEFAULT OPTIONS
github-token The GITHUB_TOKEN secret. You can use PAT if you want. string ${{ github.token }}
wait-seconds-before-first-polling Wait this interval before first polling number 10
min-interval-seconds Wait this interval or the multiplied value (and jitter) number 15
retry-method How to wait for next polling string equal_intervals exponential_backoff, equal_intervals
early-exit Stop rest pollings if faced at least 1 bad condition bool true
attempt-limits Stop rest pollings if reached to this limit number 1000
wait-list Wait only these jobs string []
skip-list Wait except these jobs string []
skip-same-workflow Skip jobs defined in the same workflow which using this action bool false
dry-run Avoid requests for tests bool false

Required GITHUB_TOKEN permissions

In public repositories, they are satisfied by default

permissions:
  contents: read # Since v2
  checks: read
  actions: read

outputs.<output_id>

  • dump
    A file path for collected resources which keeps fields than logged.
    This data is only provided for debugging purposes, so the schema is not defined.

Examples

I'm using this action for auto-merging bot PRs and wait for deploy.
See the docs for further detail.

Deadlocks

If you use this action in multiple jobs on the same repository, you should avoid deadlocks.
The skip-list, wait-list and skip-same-workflow options cover this use case.

If you changed job name from the default, you should set skip-list or roughly use skip-same-workflow

jobs:
  your_job: # This will be used default job name if you not specify below "name" field
    name: 'Changed at here'
    runs-on: ubuntu-24.04
    steps:
      - uses: kachick/[email protected]
        with:
          skip-list: |
            [
              {
                "workflowFile": "this_file_name.yml",
                "jobName": "Changed at here"
              }
            ]
        timeout-minutes: 15

Similar problems should be considered in matrix use. Because of GitHub does not provide the context.

However you can set prefix for jobMatchMode to create small skip-list to avoid this problem.

jobs:
  your_job:
    strategy:
      matrix:
        os:
          - ubuntu-24.04
          - ubuntu-22.04
    runs-on: ${{ matrix.os }}
    steps:
      - uses: kachick/[email protected]
        with:
          skip-list: |
            [
              {
                "workflowFile": "this_file_name.yml",
                "jobMatchMode": "prefix",
                "jobName": "${{ github.job }}"
              }
            ]
      - run: make test

Startup grace period

Judge whether the checkRun state at the moment.
When some jobs are triggered late after this action, we need to use the following configurations.

An example of using a wait-list.

with:
  wait-list: |
    [
      {
        "workflowFile": "might_be_triggered_after_0-4_minutes.yml",
        "optional": false,
        "startupGracePeriod": { "minutes": 5 }
      }
    ]

This action starts immediately but ignores the job missing in the first 5 minutes.

  • No need to extend wait-seconds-before-first-polling
  • Disable optional, because it is needed to check
  • Set enough value for startupGracePeriod for this purpose.
    It should be parsible with TC39 - Temporal.Duration
    e.g
    • "PT3M42S" # ISO 8601 duration format
    • { "minutes": 3, "seconds": 42 } # key-value for each unit

If not using wait-list, this pattern should be considered in your wait-seconds-before-first-polling.

Alternative candidates

gh commands, such as gh pr checks and gh run watch, should be useful if your requirements are simple.

Limitations

  • If any workflow starts many jobs as 100+, this action does not support it.
    Because of nested paging in GraphQL makes complex. See related docs for further detail.

License

The scripts and documentation in this project are released under the MIT License