No check-in yet
The monitor has not received its first ping.
No signalOne ping, one clear deadline
Heartbeat monitoring for jobs that should never fail silently.
Give every cron job, scheduled worker, backup, and batch pipeline a unique ping URL. PageLantern watches the expected schedule and turns a missed check-in into an incident with evidence.
Check-in late
The expected interval plus the configured grace window has elapsed.
MissedCheck-in restored
A new ping arrives after the missed window.
RecoveredSchedule and evidence
A small check-in with enough context to investigate.
The monitor defines when a job should report, then keeps the evidence around missed windows and recovery in one place.
Monitor contract
Define what “on schedule” means.
- Unique ping URL for each job
- Expected check-in interval
- Configurable grace window
Run evidence
See when the job went quiet.
- Last-seen and missed-window history
- Optional reported job duration
- Incident and recovery records
Location model: Heartbeat checks are hub-local schedule signals. Multi-location HTTP monitoring is a separate capability.
Example result
The job missed its window before customers noticed.
Illustrative values show the schedule, grace period, last signal, and incident outcome together.
- Job
- nightly-backup
- Expected every
- 24 hours
- Grace
- 30 minutes
- Last seen
- 25h 04m ago
- Reported duration
- 11m 42s
- Outcome
- Missed check-in
Next step
Turn one outbound ping into a useful deadline.
FAQ
Frequently asked questions
How does heartbeat monitoring catch a job that never runs?
Each job calls a unique ping URL on success. PageLantern tracks the expected interval, and when a check-in is late past its grace window it opens an incident and notifies responders — no error response required.
Do I need to install an agent on my server?
No. A single outbound request from the cron job, worker, backup, or pipeline is all that is needed; there is nothing to install on the host.
Can I see when a scheduled job last ran?
Yes. PageLantern keeps last-seen times and missed-run windows so you can prove when a scheduled task stopped and when it recovered.