Unified Observability: Monitoring Postgres Anywhere with OpenTelemetry
summary
Shows a working pipeline that collects Postgres metrics and logs from Kubernetes, bare metal and managed services using open telemetry collector with receivers (Prometheus, hostmetrics, SQL query, file log), processors (filter, attributes, batch) and exporters to Prometheus, Loki, Grafana. Demo deploys Cloud Native PG operator, collector, Prometheus and Loki; pgbbench raises connections from 3 to 14 and a crafted syntax error appears in Loki, proving end-to-end collection and correlation. Tracing is planned but not yet implemented.
AI-written summary · about
chapters (12)
- 0:00Introduction to unified observability
- 3:16Observability fundamentals and pillars
- 8:12Monitoring PostgreSQL metrics and logs
- 10:01OpenTelemetry as the unifying framework
- 13:15OpenTelemetry Collector deep dive
- 18:36Cloud Native PostgreSQL overview
- 24:03Live demo setup and configuration
- 29:13Demo: Monitoring PostgreSQL in Kubernetes
- 32:27Demo: Monitoring local PostgreSQL instance
- 37:56Unified observability with Grafana
- 39:23Future of observability with AI agents
- 41:52Q&A session
related talks
53mThe Future of Observability in Postgres
35mOpenTelemetry for Busy People
49mProfessional PostgreSQL monitoring made easy
22mEnhancing Database Observability With OpenTelemetry
50mDemystifying Kubernetes for Postgres DBAs; A Guide to Operators
22mUnified Observability
28mImplementing Observability in Kubernetes
55mCloud-Native Observability with OpenTelemetry