Hi all, I have long commits on a production database. To know why this happens I organized a snapshot system on table pg_stat_activity. When I check the result I have this row in it: postgres=# select * from public.fdj_ms_slow_stmts where pid = 2285947 \gx -[ RECORD 1 ]----+----------------------------------------------------------------------------------------------------------------------- datname | some_db pid | 2285947 leader_pid | usename | some_user application_name | PostgreSQL JDBC Driver backend_start | 2025-01-21 04:25:17.254477+01 xact_start | 2025-01-21 05:55:19.426945+01 query_start | 2025-01-21 05:55:19.428651+01 stmt_duration | 00:00:02.140691 tx_idle_duration | 00:00:00 ts_now | 2025-01-21 05:55:21.569342+01 state | active wait_event_type | Lock wait_event | advisory query_id | query | COMMIT … Is this possible ? I am really surprised ! Michel SALAIS |
