Understanding PostgreSQL Workloads: Beyond Slow Query Logs

Slow query logs have long been the starting point for PostgreSQL performance troubleshooting, but they provide only a partial view of database activity. Modern PostgreSQL deployments offer richer observability data that enables us to understand workloads, identify trends, and detect regressions before they become incidents.

In this session, we will explore the evolution of PostgreSQL query analytics, from slow query logs to pg_stat_statements, pg_stat_monitor, query fingerprints, and workload characterization. Through practical examples, we will demonstrate how to analyze workload changes, identify performance regressions after deployments, and distinguish isolated slow queries from systemic issues.

Attendees will leave with a practical framework for understanding PostgreSQL workloads and turning query analytics into actionable performance improvements.