Route alerts. Reduce the noise.
Group related alerts on a key you choose, filter out duplicates, and notify the team responsible. A group closes itself once the signal it came from has recovered.
Turn a burst of alerts into one page
A route sits between a signal crossing its threshold and a notification going out. Here is one burst of 241 alerts, from the first match to auto-resolution.
Build and maintain routing rules
Match on the fields in the alert payload, control how alerts group and how often they can notify, and keep a reviewable history of every change.
Group on a key you choose from the alert payload, and correct the grouping from the notification itself.
dedup-keyGive each source its own notification budget so one noisy sender cannot crowd out the rest.
shardsClose a group when the underlying signal has been healthy long enough, and record that it happened.
auto-resolveMatch on severity, service, environment, region or any label in the payload.
matchEvery edit is a revision with an author and a diff, so changes can be reviewed later.
revisionsRoll a route back to an earlier revision in one step.
restorePoint your existing Prometheus alerting rules at a compatible endpoint and route them here.
alertmanagerTest a route before you rely on it.
Send a sample payload through a route and see which conditions matched, which group it joined, and who would have been notified — without sending anything.
- Dry-run any payload. Paste a real alert body and watch the match happen field by field.
- See the group it joins. Including whether it would have been suppressed as a duplicate.
- See who it would wake. Resolved against the live rota, at the hour you choose to test.
- Silence with an expiry. Maintenance windows end on their own, so nothing stays muted by accident.
Quiet down your alerts.
Point your alert sources at a route, choose how they group, and decide who gets notified. Start on the free plan with no card.