LastPing ping action

Report a GitHub Actions job's start and its outcome to a LastPing monitor. LastPing raises an incident when a run fails, hangs past its expected duration, or never runs at all, and tells you on email, Slack, Discord, Telegram, Microsoft Teams and more.

The action is plain Node with no dependencies, so it works the same on Ubuntu, macOS and Windows runners whether or not curl is installed. A ping that cannot be delivered logs a warning and never fails your job, unless you ask it to.

Have admin rights on the repository? LastPing's workflow_run webhook setup is the recommended option: one signed webhook in the repository settings, no YAML changes, and it reports the name of the step that failed. Use this action when you cannot add a webhook, or when you want one monitor for one specific job.

Setup in 3 steps

1. In LastPing, create a monitor for the job and copy its ping URL (https://ping.lastping.dev/<monitor-id>).
2. In your repository, add it as a secret: Settings → Secrets and variables → Actions → New repository secret, named LASTPING_PING_URL.
3. Add two steps to the job: one at the start, and one at the end with if: always() so it runs whatever happened.

jobs:
  nightly:
    runs-on: ubuntu-latest
    steps:
      - uses: lastping-dev/ping-action@v1
        with:
          ping-url: ${{ secrets.LASTPING_PING_URL }}
          status: start

      # ... the job's steps ...

      - uses: lastping-dev/ping-action@v1
        if: always()
        with:
          ping-url: ${{ secrets.LASTPING_PING_URL }}
          status: ${{ job.status }}

Inputs

Input	Default	Description
ping-url	(required)	The monitor's ping URL. Keep it in a secret: anyone with it can ping your monitor. The action masks it in the log and never prints it.
status	success	start, success, failure, cancelled, log, or an exit code 0-255. Pass ${{ job.status }} in the final step.
body	empty	Text stored with the ping, for example a failure message. Cut to 10,000 bytes, the most LastPing stores.
run-id	<run id>-<run attempt>	Sent as ?rid=. It pairs a start with its finish, so LastPing records the run's duration.
on-cancelled	fail	What a cancelled job sends: fail alerts, log attaches a note without changing the monitor's state, skip sends nothing.
fail-on-error	false	Fail the step when the ping cannot be delivered after retries.

What each status sends

The action uses LastPing's ping API. Every request is a POST, with the body if there is one.

status	URL	Meaning in LastPing
start	<ping-url>/start	Run started; arms overrun detection
success	<ping-url>	Success
failure	<ping-url>/fail	Failure; alerts immediately
cancelled	<ping-url>/fail (see on-cancelled)	Failure, with the body GitHub Actions job cancelled unless you set one
log	<ping-url>/log	A note on the monitor; no state change
0-255	<ping-url>/<code>	0 is success, anything else a failure
skipped	nothing

A cancelled job counts as a failure by default because the work did not finish. If cancelling is routine for the job (for example with concurrency: cancel-in-progress), use on-cancelled: skip or log.

Each ping is tried up to 3 times with a 10-second timeout per attempt. Server errors, timeouts, 408 and 429 are retried; other 4xx answers are not.

Examples

Attach the reason to a failure

      - name: Tests
        shell: bash
        run: make test 2>&1 | tee test.log

      - name: Last lines of the log
        id: tail
        if: failure()
        shell: bash
        run: |
          {
            echo 'text<<EOF'
            tail -n 40 test.log
            echo EOF
          } >> "$GITHUB_OUTPUT"

      - uses: lastping-dev/ping-action@v1
        if: always()
        with:
          ping-url: ${{ secrets.LASTPING_PING_URL }}
          status: ${{ job.status }}
          body: ${{ steps.tail.outputs.text }}

A scheduled workflow

A schedule that silently stops (GitHub disables scheduled workflows in public repositories after 60 days without activity) sends no ping at all, and LastPing alerts on the missing run.

on:
  schedule:
    - cron: '0 3 * * *'

jobs:
  backup:
    runs-on: ubuntu-latest
    steps:
      - uses: lastping-dev/ping-action@v1
        with:
          ping-url: ${{ secrets.LASTPING_PING_URL }}
          status: start
      - run: ./backup.sh
      - uses: lastping-dev/ping-action@v1
        if: always()
        with:
          ping-url: ${{ secrets.LASTPING_PING_URL }}
          status: ${{ job.status }}

A matrix

Matrix jobs share a run id, so give each leg its own run-id (and, usually, its own monitor):

      - uses: lastping-dev/ping-action@v1
        with:
          ping-url: ${{ secrets.LASTPING_PING_URL }}
          status: start
          run-id: ${{ github.run_id }}-${{ github.run_attempt }}-${{ strategy.job-index }}

Use the same run-id in the final step.

Make a ping failure fail the job

      - uses: lastping-dev/ping-action@v1
        with:
          ping-url: ${{ secrets.LASTPING_PING_URL }}
          status: start
          fail-on-error: true

License

MIT