Professional PostgreSQL monitoring made easy
summary
A ready-to-run Go binary collects SQL-defined Postgres metrics with one-minute setup, noninvasive queries (no superuser objects required) and autodiscovery for Patroni clusters; metrics live in PostgreSQL or TimescaleDB and Grafana dashboards are provisioned by files. PGWatch2 ships 91 metrics with presets (exhaustive/full ≈60); a test of 45M rows shows storage falling from ~70GB to ~8GB when using Timescale compression. PGWatch3 adds parallel metric storages, tighter JSONB schema to cut sizes further, Docker Compose deployment and no built-in alerting.
AI-written summary · about
chapters (8)
- 0:00Introduction and context
- 1:50Monitoring approaches
- 6:38Postgres configuration
- 9:28Postgres tools landscape
- 13:00PG Watch 2 overview
- 17:30Dashboarding and demo
- 30:19PG Watch 3 release features
- 40:45Demo and Q&A
related talks
42mProfessional PostgreSQL monitoring made easy
50mHow To Improve PostgreSQL Monitoring and Observability With PMM - Part 2
51mHow to improve PostgreSQL Monitoring and ...
51mCollectd and Graphite: Monitoring PostgreSQL w/Style
53mThe Future of Observability in Postgres
50mHow To Improve PostgreSQL Monitoring and Observability With PMM
49mUnified Observability: Monitoring Postgres Anywhere with OpenTelemetry
25mpg_stat_monitor: A cool extension for better monitoring using PMM