DEV Community

Joao Thomazinho
Joao Thomazinho

Posted on • Originally published at cronobserver.com

Heartbeat monitoring for Kubernetes CronJobs

#kubernetes #cron #devops #observability

Kubernetes CronJobs are great until they quietly fail. Maybe the pod image can't pull, or the job exceeds backoff limits. CronObserver adds a simple heartbeat so you know when jobs don't run on time.

Why use heartbeats on CronJobs?

  • CronJobs can be suspended/deleted during deploys.
  • Out-of-resource errors happen silently.
  • Logs rotate quickly; by the time you investigate, the evidence is gone.

A single HTTP check-in after the pod completes is enough to warn you.

Step 1 – Create an observer

Define schedule/grace to match the CronJob.

Schedule: */30 * * * *
Grace: 5 minutes
Alerts: Email + webhook (Slack/Teams)
Enter fullscreen mode Exit fullscreen mode

Step 2 – Store the URL in a secret 

apiVersion: v1
kind: Secret
metadata:
  name: cronobserver-checkin
stringData:
  url: https://cronobserver.com/checkin/<token>
Enter fullscreen mode Exit fullscreen mode

Mount the secret as an env var:

env:
  - name: CRONOBSERVER_CHECKIN_URL
    valueFrom:
      secretKeyRef:
        name: cronobserver-checkin
        key: url
Enter fullscreen mode Exit fullscreen mode

Step 3 – Curl after the job completes 

command: ["sh", "-c", "run-task && curl -fsS -X POST $CRONOBSERVER_CHECKIN_URL"]
Enter fullscreen mode Exit fullscreen mode

Alternative: wrap your script to post success/failure status in JSON.

command: [
  "sh",
  "-c",
  "run-task && curl -fsS -X POST $CRONOBSERVER_CHECKIN_URL -H 'Content-Type: application/json' -d '{\"status\":\"success\"}'"
]
Enter fullscreen mode Exit fullscreen mode

Optional – Synthetic checks

Instead of waiting for a job to ping, CronObserver can proactively ping your endpoint.

synthetic_check:
  type: httpGet
  url: https://api.example.com/cron/health
  expected_status: 200
  timeout_seconds: 10
Enter fullscreen mode Exit fullscreen mode

Great for queue processors or serverless schedulers.

Alerting

  • Email or webhook per observer.
  • Webhook works with Slack, Teams, Zapier.
  • Grace periods and cooldowns reduce noise.

Links

👉 Monitor your first CronJob free: cronobserver.com/sign-up

Top comments (0)