Hey everyone, we’ve been working on pgpulse (https://pgpulse.io), a Supabase-native PostgreSQL observability product, and I wanted to share one part of the thinking behind it and get feedback from people actually running apps on Supabase. A lot of tools are good at showing metrics and alerts, but when something goes wrong, the hard part is often investigation. Not necessarily fixing it. Not collecting more data. But reducing the time it takes to understand what is actually happening. That’s the problem we’ve been focusing on as Mean Time to Investigate. We started modeling Postgres health across 11 domains: Freeze Risk Replication & Recovery Connection Pressure Lock Contention Bloat Vacuum Engine Query Throughput WAL Pipeline Disk Vitals Object Integrity Memory Fit The idea is to avoid treating database health as a flat wall of metrics. Some signals are performance issues, some are operational drift, and some are high-risk conditions that should immediately change how you investigate the system.…