Postgres monitoring tooling
The Monitoring section of awesome-postgres, all twenty entries, as catalogued on 2026-08-20 — the densest section of the catalog after GUI and Utilities. Recorded at the catalog's word (T3) except where a page exists; the substrate under most of them is written up separately as pg-stat-statements.
The section sorts cleanly by what it reads and who it reports to, and that sorting is more informative than any individual entry.
Prometheus / Grafana exporters
| Entry | Catalogued as |
|---|---|
| postgres_exporter | Prometheus exporter for server metrics |
| pg_exporter | Customizable Prometheus exporter |
| pgwatch2 | Flexible metrics monitor focusing on Grafana dashboards |
| pgMonitor | Metrics collection and visualization |
| postgresql-metrics | Extract and provide database metrics |
Nagios / Zabbix-era check plugins
| Entry | Catalogued as |
|---|---|
| check_pgactivity | Nagios plugin for monitoring clusters |
| Check_postgres | Nagios check_postgres plugin |
| libzbxpgsql | Comprehensive monitoring module for Zabbix |
| opm.io | Open PostgreSQL Monitoring suite |
Commercial SaaS
| Entry | Catalogued as |
|---|---|
| Datadog | SaaS monitoring with metrics and alerts |
| okmeter.io | Commercial SaaS with a detailed Postgres plugin |
| Instrumental | Real-time performance monitoring |
| myDBA | Performance monitoring with automated health checks |
| PMM | Percona Monitoring and Management platform |
| coroot | Open-source APM/observability, positioned as a DataDog alternative |
Point-in-time inspection, run by a human
| Entry | Catalogued as |
|---|---|
| pgmetrics | Collects server information and statistics |
| pg_view | Command-line tool showing system and per-partition stats |
| Pome | PostgreSQL metrics dashboard |
| pg_ash | Active session history by sampling pg_stat_activity |
| StatsMgr | Advanced statistics management extension |
Two entries that are not monitoring
The catalog files pgbench and dexter here, and neither belongs. pgbench is a load generator — paged separately, because the spoke's standing "no benchmark anywhere" caveat needed the instrument. dexter, "automatic indexer detecting slow queries," is an optimization tool and belongs beside hypopg: same question (which index should exist), opposite method — dexter proposes from observed slow queries, HypoPG tests a proposal against the planner.
The thing worth taking from this section
Nearly every entry above is a presentation layer over the same two sources: pg_stat_activity
for what is running now, and pg_stat_statements for what has run cumulatively. The differentiation
between twenty projects is transport (Prometheus, Nagios, SaaS agent) and rendering, not measurement.
That is why the spoke paged pg-stat-statements from the PostgreSQL documentation instead of paging
five tools that wrap it — and why pg_ash, which samples pg_stat_activity to build an Oracle-style
active session history, is the entry here doing something the others aren't.
Next candidates
pg_ash (sampled session history as a distinct technique) and PMM or pganalyze for how a commercial product frames the same data. Both are lower priority than the Optimization section, which aims at an open growth edge.
Related
pg-stat-statements · pgbench · hypopg · pev2 · query-execution-plan · postgres-query-optimization-tooling · awesome-postgres · operational-databases