Professional PostgreSQL monitoring made easy
summary
A practical walkthrough of pgwatch shows a single static Go binary that collects almost every stats collector table as a plain SQL metric, ships to TimescaleDB compression or Prometheus/Grafana, and exposes a simple web UI with autodiscovery and presets. It covers ~72 default metrics, supports Postgres 11–17, and estimates six databases for three months produce 45 million rows = 70 GB raw (compressing to ~8 GB with TimescaleDB). Dashboards are templates; alerts use Grafana/Prometheus.
AI-written summary · about
chapters (14)
- 0:00Introduction to Cybertec
- 1:37Importance of monitoring
- 3:01Levels of monitoring
- 4:38PostgreSQL data analysis
- 7:34Tools and extensions
- 9:37Real-time monitoring tools
- 12:33Continuous monitoring options
- 14:35Introduction to pgwatch
- 17:40Architecture and configuration
- 19:24Features and storage
- 22:49Deployment and setup
- 24:51Dashboards and visualization
- 29:08Web UI and customization
- 36:15Q&A and conclusion
related talks
49mProfessional PostgreSQL monitoring made easy
51mCollectd and Graphite: Monitoring PostgreSQL w/Style
51mHow to improve PostgreSQL Monitoring and ...
50mHow To Improve PostgreSQL Monitoring and Observability With PMM - Part 2
50mHow To Improve PostgreSQL Monitoring and Observability With PMM
53mThe Future of Observability in Postgres
49mUnified Observability: Monitoring Postgres Anywhere with OpenTelemetry
25mPoint-in-time query tuning and observability with pg_stat_statements