Error in percona logs - pg_stat_monitor: pg: out of memory

PostgreSQL
1

in the pmm-agent log, we get this error: “pmm-agent: time=“2025-08-12T04:08:25.731+00:00” level=error msg=“getNewBuckets failed: failed to query pg_stat_monitor: pq: out of memory” AgentID=daf6b1e1-8cec-4f84-a85b-e6d2f3d346f1 component=agent-builtin type=qan_postgresql_pgstatmonitor_agent”

huge_pages is enabled in postgresql. the system has huge pages configured based on postgresql memory usage.

when the agent issues this error, metrics from PostgreSQL to Percona stop being received.

pg_stat_monitor settings (2.2.0): name | setting ----------------------------------------------±-------- pg_stat_monitor.pgsm_bucket_time | 60 pg_stat_monitor.pgsm_enable_overflow | about pg_stat_monitor.pgsm_enable_pgsm_query_id | about pg_stat_monitor.pgsm_enable_query_plan | off pg_stat_monitor.pgsm_extract_comments | off pg_stat_monitor.pgsm_histogram_buckets | 20 pg_stat_monitor.pgsm_histogram_max | 30000 pg_stat_monitor.pgsm_histogram_min | 50 pg_stat_monitor.pgsm_max | 256 pg_stat_monitor.pgsm_max_buckets | 10 pg_stat_monitor.pgsm_normalized_query | pg_stat_monitor is enabled .pgsm_overflow_target | 1 pg_stat_monitor.pgsm_query_max_len | 2048 pg_stat_monitor.pgsm_query_shared_buffer | 20 pg_stat_monitor.pgsm_track | Top application names pg_stat_monitor.pgsm_track_ | disabled pg_stat_monitor.pgsm_track_planning | disabled pg_stat_monitor.pgsm_track_utility / enabled

pmm-agent version 3.3.0

pg_stat_monitor version 2.2.0

2

Hi @ANDiv1976,

Have you checked the PostgreSQL and system/OS logs also? Have you checked the memory graphs for that server under PMM?

3

The problem of running out of memory is due to the fact that pg_stat_monitor does not collapse the same type of queries.