Schedule - Nordic PGDay 2026
Actionable Observability: Finding the root cause of Slow Queries in Postgres
Date: 2026-03-24
Time: 17:00–17:50
Room: Vision 2+3
Level: Intermediate
Postgres offers many tools for query observability, but it’s a struggle to use them effectively. Either they collect too little data and miss performance problems or log everything creating noise and confusion. In this talk, we present a practical framework to configure Postgres logging and query analysis with a single goal: identify the queries worth investigating. Drawing on experience monitoring Postgres across managed services, Kubernetes, and self-hosted environments, we explore how to balance visibility, performance impact, and signal-to-noise. Learn how tools like pg_stat_statements, slow query logging, auto_explain, and built-in metrics complement each other—and where they fall short when used alone. We also look at common observability mistakes, such as premature global tuning, over-instrumentation, and logging without a clear question. Finally, we show how pganalyze supports these practices and how observability data paired with AI tools can enhance investigation workflows.
