Database Monitoring
Observe query availability and connectivity for critical database-backed workflows, and alert on timeouts, auth failures or network issues. Database monitoring confirms that the data stores your application depends on are not just reachable, but able to answer queries within an acceptable time.
What it checks
A database monitor connects to a database on a schedule, runs a validation query and reports the outcome:
- Connectivity and auth - whether the monitor can reach the database and authenticate successfully over the network path it uses.
- Query health - whether a validation query completes successfully, and an alert when it times out or returns an error.
- Performance - how long queries take to complete, so degradation is visible before it causes an outage.
- Resilience - read replica availability and failover readiness for setups that depend on them.
Database monitoring supports common engines including Postgres, MySQL, Mongo and Redis.
Configuration
Database monitors are configured around the connection and the query used to test it:
- Connection details - the host, port and database name needed to reach the instance, along with the engine type.
- Credentials - the username and password the monitor uses to connect. Credentials are handled securely and supplied per monitor; use a least-privilege account scoped to the validation query rather than an administrative login.
- Validation query - the query the monitor runs to confirm the database is responding, typically a lightweight read.
- Check interval - how often the monitor connects and runs the query.
- Timeout - how long to wait for the connection and query before the check is treated as failed.
Keep the validation query lightweight and read-only - something like a trivial select. A heavy query adds load to the database every interval and makes it harder to tell a real outage apart from a slow query.