CNPGClusterOffline¶
| Severity | critical |
| Fires when | No database instance is running, for 5 minutes |
What it means¶
No database instance of your PostgreSQL service has run for 5 minutes. The alert also fires when the platform cannot read the metrics of any database instance.
What happens if you do nothing¶
The database is not available. Your applications cannot connect to it.
What to do¶
- Check if you paused the instance. A paused instance has no running database, so this alert is expected. Resume the instance, or mute the alert while the instance is paused.
- Check the status of the instance and its pods. Open the Technical Details on the instance detail page. The Pods table shows the Phase and Ready state of each pod. See Monitor an Instance.
- Check the logs of the pods for errors. See View Logs.
- Check the common causes:
- Full storage: the Persistent Volumes table shows a volume at or near 100%. See PersistentVolumeFillingUp.
- Out of memory: the logs or the pod status show that the database was stopped because it used too much memory. See MemoryCritical.
- If you cannot find the cause, contact support at once. Include the instance name and the alert name.