🎟️
turnstyle

A GitHub Action for serializing workflow runs

🤔 why bother

GitHub Actions is an event-oriented system. Your workflows run in response to events and are triggered independently and without coordination. In a shared repository, if two or more people merge pull requests, each will trigger workflows without regard to one another.

This can be problematic for workflows used as part of a continuous deployment process. You might want to let an in-flight deployment complete before progressing further with the next workflow. This is the usecase turnstyle action targets.

🤸 Usage

The typical setup for turnstyle involves adding job step using softprops/turnstyle@v1.

name: Main

on: push

jobs:
  main:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout
        uses: actions/checkout@v2
+     - name: Turnstyle
+       uses: softprops/turnstyle@v1
+       env:
+         GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
      - name: Deploy
        run: sleep 30

Timing out

To avoid waiting prolonged periods of time, you may wish to bail on a run or continuing a workflow run regardless of the status of the previous run.

You can bail from a run using the built-in GitHub Actions jobs.<job_id>.timeout-minutes setting

name: Main

on: push

jobs:
  main:
+   timeout-minutes: 5
    runs-on: ubuntu-latest
    steps:
      - name: Checkout
        uses: actions/checkout@v2
      - name: Turnstyle
        uses: softprops/turnstyle@v1
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
      - name: Deploy
        run: sleep 30

You can also limit how long you're willing to wait before moving on with jobs.<job_id>.steps.with.continue-after-seconds

name: Main

on: push

jobs:
  main:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout
        uses: actions/checkout@v2
      - name: Turnstyle
        uses: softprops/turnstyle@v1
        with:
+         continue-after-seconds: 180
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
      - name: Deploy
        run: sleep 30

or before aborting the step with jobs.<job_id>.steps.with.abort-after-seconds

name: Main

on: push

jobs:
  main:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout
        uses: actions/checkout@v2
      - name: Turnstyle
        uses: softprops/turnstyle@v1
        with:
+         abort-after-seconds: 180
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
      - name: Deploy
        run: sleep 30

Finally, you can use the force_continued output to skip only a subset of steps by setting continue-after-seconds and conditioning future steps with if: ! steps.<step id>.outputs.force_continued

name: Main

on: push

jobs:
  main:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout
        uses: actions/checkout@v2
      - name: Turnstyle
        id: turnstyle
        uses: softprops/turnstyle@v1
        with:
+         continue-after-seconds: 180
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
      - name: Deploy
+       if: ! steps.turnstyle.outputs.force_continued
        run: sleep 30

inputs

Name	Type	Description
continue-after-seconds	number	Maximum number of seconds to wait before moving forward (unbound by default). Mutually exclusive with abort-after-seconds
abort-after-seconds	number	Maximum number of seconds to wait before aborting the job (unbound by default). Mutually exclusive with continue-after-seconds
poll-interval-seconds	number	Number of seconds to wait in between checks for previous run completion (defaults to 60)
same-branch-only	boolean	Only wait on other runs from the same branch (defaults to true)

outputs

Name	Type	Description
force_continued	boolean	True if continue-after-seconds is used and the step using turnstyle continued. False otherwise.

environment variables

The following are required as step.env keys

Name	Description
GITHUB_TOKEN	GITHUB_TOKEN as provided by secrets

cost of coordination

At this time there is no way to coordinate between workflow runs beyond waiting. For those using private repositories, you are charged based on the time your workflow spends running. Waiting within one workflow run for another to complete will incur the cost of the time spent waiting.

Doug Tangren (softprops) 2020