Reliability
Monitoring and database logs
Use metrics, logs, and alerts to catch failures before customers do.
7 minute read ยท Updated September 2026Core signals
- Connections and saturation.
- CPU, memory, storage, and growth rate.
- Query latency and slow-query volume.
- Backup age and WAL continuity.
- Replica lag and health where replicas are enabled.
- Provisioning worker and control-plane availability.
Database logs
Use Database logs to filter operational events and PostgreSQL messages by database, severity, and time. Copy safe error details into a support ticket, but remove passwords, connection strings, customer data, and tokens.
Actionable alerts
- Every alert should name the affected resource.
- Include a runbook or next action.
- Route severity to the correct channel.
- Deduplicate repeated symptoms.
- Close the loop with a resolution event and incident record.
Was this guide useful?
Get helpIf anything is unclear or incorrect, tell the support team.