Resources
PostgreSQL Problems, with the Query to Run and the Warning You Could Have Had
Each guide takes one problem that pages a PostgreSQL DBA: what the server says, the queries that show who or what caused it, the fixes in the order that does not make it worse, and how Radar Engine reports it before the outage. The SQL runs on PostgreSQL 16 or later with a role in pg_monitor.
Space
- Transaction ID wraparound
How close each database is, which tables carry the age, what keeps vacuum from freezing, and the fix in order. - pg_wal fills the disk
The replication slot nobody reads, the archiver that fails, the setting that keeps WAL, and the bound that prevents it.
Load
- Idle in transaction: find the session that blocks
The root of a lock queue, the ALTER TABLE everybody waits behind, and the timeouts that prevent it. - Sorry, too many clients already
How close the connection limit is, who holds the connections, and why raising max_connections comes last.
Replication
- Replication lag
Whether the WAL waits to be sent, written or replayed, the cause behind each stage, and the fix in order.
Watch every instance for these
Scops runs 29 Radar detectors on every PostgreSQL instance and reports each of these problems while there is still time to act, with no agent on the server and one role in pg_monitor. See PostgreSQL monitoring with Scops.