Monitor GitHub Actions Scheduled Workflows

GitHub runs scheduled workflows on a best-effort basis. Its docs say a scheduled run “can be delayed during periods of high loads” and that “some queued jobs may be dropped”. In public repositories, GitHub disables scheduled workflows after 60 days without repository activity. GitHub doesn’t tell you when it drops a run or disables a workflow. A Chronos monitor alerts on a failed run and on a run that never happened.

Add the ping step

1. Create a monitor with the same cron expression and timezone as the workflow’s schedule. GitHub uses UTC unless the schedule sets timezone.
2. In the repository’s Settings, open Secrets and variables, then Actions, and add a repository secret named CHRONOS_PING_URL with the monitor’s ping URL.
3. Add a final step to the job:

.github/workflows/nightly-report.yml

name: Nightly report

on:
  schedule:
    - cron: '17 6 * * *'
  workflow_dispatch:

jobs:
  report:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v7
      - run: ./scripts/report.sh

      - name: Ping Chronos
        if: always()
        env:
          PING_URL: ${{ secrets.CHRONOS_PING_URL }}
          JOB_STATUS: ${{ job.status }}
          RUN_URL: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}
        run: |
          if [ "$JOB_STATUS" = "success" ]; then
            curl -fsS -m 10 --retry 5 -o /dev/null "$PING_URL" || true
          else
            curl -fsS -m 10 --retry 5 -o /dev/null --data-raw "$JOB_STATUS: $RUN_URL" "$PING_URL/fail" || true
          fi

if: always() runs the step even when an earlier step failed or the run was cancelled. job.status is success, failure, or cancelled. On anything but success the step sends a fail ping, and the ping’s body puts the run’s link in the down alert. || true keeps a lost ping from failing the run. curl still prints the error in the step’s log, and the monitor goes down after the grace period. The curl flags are the same as for cron.

Pick the schedule and grace period

GitHub’s high-load times include the start of every hour, so schedule off the hour, as 17 6 * * * does. Set the monitor’s grace period to cover the workflow’s run time plus the delays you see in the Actions tab. If runs regularly start 20 minutes late, a 30-minute grace period keeps the monitor quiet on a normal day.

Check it works

workflow_dispatch adds a Run workflow button in the Actions tab. Run the workflow once and the monitor moves from new to up. To test the alert, add exit 1 to a step before the ping and run the workflow again. The monitor goes down, and the alert links to the run. Remove the exit 1 and run it once more to bring the monitor back up.