Read monitoring status and missing metrics
Separate service state, freshness, absent workload and actual failure.
Before you begin
Administrator access and a fresh sample.
Use the controls offered by your installed version and account permissions. This guide describes the workflow; it does not grant additional access.
Follow the steps
- Check service/incident views in Monitoring.
- Read sample timestamps before interpreting values.
- Read explanations for Not in use and Unavailable.
- Investigate unhealthy services through workload and log evidence.
- Acknowledge or resolve incidents according to actual state and permission.
Check the result
The diagnosis identifies evidence, scope and next action.
If something goes wrong
New servers may have no tenant FPM queues or access logs. Kernel visibility can also limit cgroup and I/O telemetry; unknown must not become zero.
Prepare useful details for support