Wait Event Heatmaps: See What PostgreSQL Is Actually Waiting On

Wait Event Heatmaps: See What PostgreSQL Is Actually Waiting On If you have ever stared at a Grafana dashboard full of green CPU and memory panels while your users complain about slow queries, you have experienced the fundamental blind spot of resource utilization monitoring. PostgreSQL has a built-in profiler that most teams completely ignore — wait events. The Problem Your database feels slow but the usual metrics look fine. CPU is at 30%. Memory is stable. Disk I/O graphs show nothing dramatic. The problem is invisible because you are looking at resource utilization metrics, which tell you how much of each resource is consumed but not what PostgreSQL is spending its time doing. PostgreSQL tracks exactly what every backend is waiting on at any given moment. These are wait events, and they fall into categories: IO (reading or writing data files), Lock (heavyweight row or table locks), LWLock (lightweight internal locks like buffer mapping or WAL insertion), BufferPin (waiting for a sh