Monitoring Lifecycle
The exact specification of how checks run, when they become incidents, and when incidents resolve.
Check Intervals
Section titled “Check Intervals”Checks run on a 1-minute cycle against monitors that are due. The effective interval is determined as follows:
The effective interval is the longer of the monitor’s configured interval and the plan minimum.
| Plan | Minimum check interval |
|---|---|
| Free | 5 min |
| Pro | 3 min |
Paused monitors are not checked.
Monitor Statuses
Section titled “Monitor Statuses”| Status | Meaning |
|---|---|
up | Healthy. All check conditions are met |
down | Failing. Unexpected status code, timeout, keyword condition violation, etc. |
degraded | Indeterminate. The check infrastructure could not determine the result. The current HTTP / TCP / Ping monitors only return up / down, so this state does not occur |
paused | Paused |
unknown | Not yet checked or otherwise unknown |
Automatic Incident Creation
Section titled “Automatic Incident Creation”A single failed check never triggers a notification, to avoid false alarms.
- Each
downcheck increments a consecutive-down counter - After 5 consecutive
downchecks, an incident (ongoing) is created automatically - An incident-created notification (severity: critical) is sent at the same time
- While an unresolved incident exists, no duplicate incident is created
Grace period for new monitors
Section titled “Grace period for new monitors”A newly created monitor holds off incident detection until its first up check (to avoid false alarms from misconfiguration). However, if the monitor goes down 5 times in a row without ever being up, the grace period is cut short and an incident is created.
Recovery and Automatic Resolution
Section titled “Recovery and Automatic Resolution”- When a check returns to
up, the consecutive-down counter resets and any unresolved incidents (ongoing or acknowledged) are automatically resolved - A recovery notification is sent at the same time
- “Acknowledged” indicates that someone is handling the incident; acknowledged incidents are also auto-resolved on recovery
- Manual resolution resets the consecutive-down counter. If the monitor remains
down, a new incident is created and notifications are sent after 5 more consecutive failures
Maintenance Windows
Section titled “Maintenance Windows”Maintenance windows can be set per monitor or per service.
- During the window, incident detection is fully suspended and the consecutive-down counter is reset
- With auto-resolve enabled the window ends at its end time; otherwise it continues until manually ended