Skip to content
Feature

Cron job and heartbeat monitoring

A scheduled job can't be checked from outside, so it checks in instead. heyUptime gives each job a secret ping URL; the job requests it when it finishes. If no check-in arrives within the expected period plus a grace time, the job counts as down and you get an alert, exactly like a website.

How a heartbeat works

In short: The job calls the URL; heyUptime watches the clock. Silence past the deadline is an incident.

  • Add a heartbeat with a name, how often the job runs (every minute up to every week) and how late is acceptable (1 minute up to 1 hour).
  • Append a request to the ping URL at the end of the job, for example curl -fsS https://heyuptime.com/ping/… >/dev/null.
  • Each check-in sets the next deadline to now plus the period plus the grace time.
  • If the deadline passes with no check-in, an incident opens and alerts go out. The next check-in closes it with a recovery message.

Periods and grace times

Pick the period that matches the job's schedule and a grace time that covers a slow run. A nightly backup that usually takes 20 minutes might use a period of 1 day and a grace of 30 minutes.

SettingOptions
Period (how often the job runs)1 minute, 5, 15, 30 minutes, 1 hour, 6, 12 hours, 1 day, 1 week
Grace (how late is still fine)1, 5, 15, 30 minutes, 1 hour
DefaultEvery day, 15 minutes of grace

What the alert says

A missed check-in is reported with the same words on every channel: "No check-in for 1 day 15 minutes. It was expected every 1 day." A job that stays dead is counted once per period, not every second, so one broken cron entry produces one open incident, not a flood.

Details worth knowing

  • A new heartbeat is not armed until its first check-in, so adding one never raises an alarm before the job has had a chance to run.
  • Changing the period or grace measures the new schedule from the last check-in; it never triggers an alert by itself.
  • Check-ins closer than 5 seconds apart are acknowledged but not stored, so a script stuck in a loop can't fill your history.
  • Pausing a heartbeat stops the clock; check-ins are answered but not counted.
  • Heartbeats count towards your plan's monitor limit (10, 50 or 200) and can appear on status pages.

Try it before you sign up

The cron job tester at /tools/heartbeat-tester makes a temporary URL. Have your job call it and watch the call arrive, so you know the command works before you rely on it. The cron schedule pages explain the common expressions.

Questions

What is a heartbeat monitor?

A monitor that expects your job to call a URL on a schedule. heyUptime can't run your backup, but it can tell when the backup didn't run.

Can I use this for things other than cron?

Yes. Anything that runs on a schedule and can make an HTTP request: Windows scheduled tasks, Kubernetes CronJobs, GitHub Actions on a timer, a Raspberry Pi script, a nightly data import.

What if the job fails halfway?

Put the ping at the very end, after the real work. Then a job that crashes never checks in and is reported as down. With curl -f, a failed ping also fails the job's exit code.

Does a heartbeat count as a monitor?

Yes. Heartbeats and websites share the plan's limit: 10 on Starter, 50 on Pro, 200 on Business.

Find out the night the backup doesn't run.

Add a heartbeat, paste one line into your script, and heyUptime tells you when the job goes quiet. Free for 14 days, no card.

Start watching free

Updated October 8, 2026.