Collectd and Graphite: Monitoring PostgreSQL w/Style
summary
Shows a working collectd + Graphite stack for PostgreSQL monitoring with concrete configs not vague promises it uses the collectd postgres plugin to run arbitrary queries, polls every 10 seconds, and ships through carbon into whisper with configurable retention so you know storage cost; demonstrates replication lag and idle‑transaction queries, a moving average and saved dashboards, and cites real load numbers like ~15,000 TPS and ~2 billion queries/day so you can judge scale and effort.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 1:27Why Graphical Monitors
- 2:08Graphite
- 2:25Why Collectd
- 3:00Plugins
- 3:23Collectd Overview
- 4:00Collectd Plugins
- 4:40Configuration Management Systems
- 5:02Graphs
- 5:30Dashboards
- 7:25Users
- 8:28Collectd Configuration Files
- 9:23Networkcom Files
- 10:29Local System
- 12:19PostgreSQL Plugin
- 12:42Query Sections
- 15:37Startup Scripts
- 16:33Installing Graphite
- 17:51Dependencies
- 21:21Configure Collectd
- 27:44TPS query
- 29:42Growth query
- 32:51Liner notes
- 34:08Demo
- 37:30Dashboard
- 43:41Graph
- 47:41Multiple Graphs
related talks
49mProfessional PostgreSQL monitoring made easy
42mProfessional PostgreSQL monitoring made easy
50mHow To Improve PostgreSQL Monitoring and Observability With PMM - Part 2
51mHow to improve PostgreSQL Monitoring and ...
53mThe Future of Observability in Postgres
50mHow To Improve PostgreSQL Monitoring and Observability With PMM
31mMaking Data Dance
46mPostgreSQL as a Graph Database: Who Grabbed a Beer Together?