Alerts¶
An alert has the status open or resolved. Acknowledgment is recorded separately (who and when). Acknowledging an open alert also sets it to resolved; auto-close resolves it when the target recovers.
When an alert is created¶
- A DOWN alert is created when a service is down longer than the Down Alert
Delay (
metrics_timeout_minutes, 1-10 min, default 5). - The
unknownstatus never alerts, for example while the quality gate pauses checks. - DOWN notifications have a 5-minute cooldown. A RECOVERY is sent immediately.
- A certificate alert is created at 7 or fewer remaining days and resolves after
renewal. It appears in the inbox as
monitoring.certificate.warning.
Fields¶
| Field | Content |
|---|---|
| Target / service | affected target and check type |
| Start / end, duration | outage timestamps and length |
| Recipients | notified recipients |
| Acknowledgment | who and when |
Auto-close can be switched per target.
Acknowledging¶
Acknowledge alerts one by one or in bulk. This is allowed for admins and users
with the alert-handler permission (can_close_alerts). See Users & roles.