Understand incidents and recovery

An incident spans a trouble episode. When evaluated health becomes degraded or down, Pingstack opens an incident or updates the existing open one. Severity changes between degraded and down remain part of that episode. Full return to up resolves it.

Transition	Incident effect
Up to degraded/down	Open trouble episode
Degraded to down, or down to degraded	Update the open episode rather than creating a second one
Degraded/down to up	Resolve the episode
No status change	No new transition merely because another sample arrived

Review the check’s incidents for opening time, recovery time, duration and cause. Dashboard incident history combines episodes for its checks and offers state, check and date filters. Incident starts and incident overlap are different counts; see history semantics.

Notification routing uses the check’s assigned groups and eligible recipients. An incident and a notification delivery are separate records; consult delivery history to diagnose missing messages.

Recovery is determined by monitoring health, not by posting an incident note. The API exposes safe incident timing and duration, but does not support acknowledging, manually resolving or writing incident updates.