Database Observability: Fleet Overview
The Database Observability: Fleet Overview dashboard shows the status and load of all monitored PostgreSQL instances. It helps identify unavailable, overloaded, and lagging instances.
Who it is for: on-call engineers, database administrators, and operations managers.
Data sources: db_metrics*, db_activity*, db_replication*, db_meta*, .smos_incident-*.
Main Sections
1. Fleet Status
Shows an overview of the infrastructure:
- number of instances, TPS, and proportion of occupied connections
- number of unavailable instances and open incidents
- availability of clusters and individual instances
2. Connections and Activity
Shows load trends and instance status:
- proportion of occupied connections
- number of active sessions
- cluster, environment, criticality, and role of each instance
- TPS, cache hits, replication lag, and overall health assessment
3. Connection Saturation and Database Versions
Compares instances by connection saturation and shows the distribution of instances by PostgreSQL version.
Links to Other Sections
| Panel | Destination |
|---|---|
Instances table | Database Observability: Instance State |
Instances in Fleet metric | Database Observability: Capacity and Growth |
Unresponsive Instances metric | Database Observability: Collection Diagnostics |
Open Incidents metric | Database Observability: Incident Manager Integration |