Monitors & detections
Give passing hiccups a little patience
Use consecutive failures and successes to confirm a status change.
Reviewed
Choose the confirmation thresholds
Expand Status confirmation. Failures before downtime and Successes before recovery each accept 1 to 10 consecutive checks. Both default to 1.
Every result remains in history. Incident transitions and alerts follow confirmed status, so a failed individual check does not always create a new incident.
An example with three failures
With Failures before downtime set to 3, the sequence failure, failure, success does not confirm downtime: the success resets the failure streak. Three consecutive failures do confirm downtime.
Incident timing starts when downtime is confirmed, not retroactively at the first failed check. With Successes before recovery set to 2, a down monitor needs two consecutive successful checks before recovery is confirmed.
What resets a streak
An opposite result resets the relevant streak. Manual checks count too. Saving check settings or pausing/resuming starts a fresh streak without rewriting the confirmed status. Maintenance also separates confirmation streaks.
A new monitor awaits enough results to confirm its status. Thresholds count results, not elapsed minutes; the check interval, duration and scheduling determine how long confirmation takes.