Database Monitoring
Connection health, replication lag, and slow queries across every engine an org has registered
Databases
48
Healthy
43
Degraded
4
Down
1
Registered databases
PostgreSQL · MySQL / MariaDB · SQL Server · Oracle · Aurora · Redshift — credentials stored per organization
No matching results
Catalog findings
Schema drift, volume drops, and redundancy — read from the catalog on every load, never stored
No matching results
Slowest queries (24h)
How database checks run
Each database an organization registers is probed on its own schedule. Credentials are held as secret references underorg_id, never in the browser.
- Connectivity— driver-level connect with a short timeout
- Replication— lag measured against the primary's LSN / GTID
- Saturation— open connections against the configured pool ceiling
- Slow queries— sampled from the engine's own statement statistics
A failed probe raises an event into the same alert pipeline as API and ETL signals, so database incidents cluster with whatever else broke at the same moment.
