Alert Runbooks¶
A runbook describes what to do when an alert fires on your instance. Each alert has its own page: what the alert means, what happens if you do nothing, and the steps to fix it.
You see the alerts in the Alerts card on the instance detail page, and by email if you have set an alert email address. See Monitoring for how alerting works.
Alerts for all services¶
These alerts are available for all services with user alerting.
| Alert | Severity | Fires when |
|---|---|---|
| MemoryCritical | critical |
Memory usage is above 85% of the limit for 2 hours |
| PersistentVolumeExpectedToFillUp | critical |
Less than 15% storage is free, and the volume will be full within 4 days at the current rate, for 1 hour |
| PersistentVolumeFillingUp | critical |
Less than 3% storage is free, for 15 minutes |
Runbooks per service¶
Each service has its own runbook pages. They add the steps for that service to the general alerts, and they cover the alerts that only this service has.
Muting an alert¶
Fix the cause first. Mute the alert only when you know about the problem and have a plan. Write the plan into the reason when you mute the alert. Other members of your organization see the reason in the audit log.
When to contact support¶
Contact support if:
- An alert fires, but the usage in the Technical Details looks normal.
- The storage size or compute plan change does not apply.
- The instance already has the largest storage size or compute plan.
- You do not know what causes the problem.
Include the instance name and the alert name in your request.