Cron monitoring examples: crontab, Docker

The principle is the same everywhere: make an HTTP request to the ping URL when the job succeeds (and, better, to /start at the beginning and /fail on failure). Replace <ping-key> with your check’s key. The -fsS -m 10 --retry 3 options are explained in the ping API.

The simplest: the ping is sent only if the script succeeds (&&).

0 2 * * *  /usr/local/bin/backup.sh && curl -fsS -m 10 --retry 3 https://app.silencewatch.com/p/<ping-key>

To measure the duration and report failure explicitly, use a small wrapper script:

#!/bin/sh
URL=https://app.silencewatch.com/p/<ping-key>
curl -fsS -m 10 --retry 3 "$URL/start"
/usr/local/bin/backup.sh
curl -fsS -m 10 --retry 3 "$URL/$?"

In the service the timer starts:

[Service]
Type=oneshot
ExecStart=/bin/sh -c '/usr/local/bin/backup.sh && curl -fsS -m 10 --retry 3 https://app.silencewatch.com/p/<ping-key>'

[Timer]
OnCalendar=*-*-* 02:00:00
Persistent=true

[Install]
WantedBy=timers.target

docker run --rm my-image /app/job.sh \
  && curl -fsS -m 10 --retry 3 https://app.silencewatch.com/p/<ping-key>

In a docker-compose.yml you can chain in the service’s command: command: sh -c "/app/job.sh && curl -fsS -m 10 --retry 3 $PING_URL", provided the image contains curl.

apiVersion: batch/v1
kind: CronJob
metadata:
  name: backup
spec:
  schedule: "0 2 * * *"
  jobTemplate:
    spec:
      template:
        spec:
          restartPolicy: OnFailure
          containers:
            - name: backup
              image: my-image:latest
              command: ["/bin/sh", "-c"]
              args:
                - /app/backup.sh && curl -fsS -m 10 --retry 3 "$PING_URL"
              env:
                - name: PING_URL
                  valueFrom:
                    secretKeyRef:
                      name: silencewatch
                      key: ping-url

The ping URL is a secret: put it in a Secret, not in the manifest.

on:
  schedule:
    - cron: "0 2 * * *"
jobs:
  nightly:
    runs-on: ubuntu-latest
    steps:
      - run: ./scripts/nightly.sh
      - name: Ping SilenceWatch
        if: success()
        run: curl -fsS -m 10 --retry 3 "${{ secrets.SILENCEWATCH_PING_URL }}"
      - name: Report the failure
        if: failure()
        run: curl -fsS -m 10 --retry 3 "${{ secrets.SILENCEWATCH_PING_URL }}/fail"

GitHub’s scheduled workflows can be delayed or skipped; that is exactly the kind of absence a dead man’s switch catches.

import requests

URL = "https://app.silencewatch.com/p/<ping-key>"

requests.get(f"{URL}/start", timeout=10)
try:
    run_job()
except Exception:
    requests.get(f"{URL}/fail", timeout=10)
    raise
else:
    requests.get(URL, timeout=10)

const PING_URL = 'https://app.silencewatch.com/p/<ping-key>';

await fetch(`${PING_URL}/start`, { signal: AbortSignal.timeout(10_000) });
try {
  await runJob();
  await fetch(PING_URL, { signal: AbortSignal.timeout(10_000) });
} catch (error) {
  await fetch(`${PING_URL}/fail`, { signal: AbortSignal.timeout(10_000) });
  throw error;
}

$url = "https://app.silencewatch.com/p/<ping-key>"
& "C:\Jobs\backup.ps1"
if ($?) { Invoke-RestMethod -Uri $url -TimeoutSec 10 }
else { Invoke-RestMethod -Uri "$url/fail" -TimeoutSec 10 }

For Spring Boot

Write nothing: the Spring Boot starter discovers @Scheduled methods and Quartz jobs and sends the pings for you.