Monitoring that explains what happened
Cron job monitoring with heartbeat alerts
A website can be online while its nightly backup silently stops. Give each important job a heartbeat monitor and signal only after the work succeeds.
Success is a deliberate signal
Create a monitor, copy its one-time ping URL and call it after a successful job. GET and POST are supported. In shell, backup.sh && curl --fail --max-time 10 "$HEARTBEAT_URL" sends a ping only when the preceding command succeeds.
A deadline with room for normal variation
Set the expected interval and a grace period. Healthy means the next heartbeat is not overdue; Late means the interval has elapsed; confirmed missed deadlines become Down and open an incident.
Recovery without guesswork
The next accepted heartbeat records recovery and resolves the incident. The detail page shows the last ping and next expected time. Rotate the secret URL if it is exposed.
Common questions
Does a heartbeat verify backup contents?
No. It confirms that your job sent a success signal. Validate backup integrity and restoreability separately before sending that signal.
What if the job never sends its first ping?
Its deadline begins when the monitor is created or resumed. A missing first heartbeat can therefore trigger an incident.