Heartbeats for backups and scheduled jobs

Backups and scheduled jobs (sending invoices, syncing orders, clearing old data) run in the background. When one stops, the website still loads, so nobody notices until something goes wrong. A heartbeat watches for that silence.

How it works

1. You add a heartbeat and say how often the job runs, for example every day.
2. We give it a private check-in address.
3. The job visits that address every time it finishes.
4. If a visit doesn't arrive on time, we alert you.

Add a heartbeat

1. Open Heartbeats in the menu.
2. Give it a name, like Nightly backup.
3. Choose how it runs:
   • Every…: every so many minutes, hours, days, weeks or months, counted from each check-in.
   • At set times: the days and times it runs, like Monday to Friday at 9:00 AM and 6:00 PM, in your time zone.
   • Cron: if the job already has a cron schedule, like 0 2 * * *, paste it in.
4. Choose how late it may be before you're alerted. Give some room: a backup that usually takes 20 minutes might take longer on a busy night.
5. Click Add heartbeat. Its page shows the check-in address.

Connect the job

The job only needs to call the check-in address when it finishes: any request works, from any language or tool. The heartbeat's page has ready-to-copy examples, with its own address filled in, for:

• Cron on a server: one line after the job's command.
• Laravel: the scheduler has it built in (pingOnSuccess and pingOnFailure).
• PHP, Node.js and Python: a call at the end of the job, and the failure address if it goes wrong.
• WordPress: a call at the end of a scheduled task.
• No-code tools and backup services: a last GET request step in Zapier, Make or n8n, or the ping URL setting of a backup plugin.
• To test it: open the address in your browser. It says OK, and the heartbeat shows the check-in.

What you'll see

• Waiting for first check-in: nothing is checked until the job checks in once.
• OK, or Running late while it's past due but still within the time you allowed.
• Missing: the check-in didn't arrive. You get a MISSED email (and webhook, Slack or Discord message), reminders if you turned them on, and a BACK email when it checks in again.

Report a failure

If the job knows it failed, have it visit the address with /fail on the end. We alert you straight away instead of waiting for the next check-in.

Tip: Heartbeat alerts go to the account owner and team members with alert emails on. Client viewers don't see heartbeats. Pause one while you change the job, so it doesn't alert.