Postgres Summit US2026

Everyone knows how to create a PostgreSQL index. Far fewer know what happens to it six months later. While the PostgreSQL community has plenty of content on index types and query planning, the operational side of index lifecycle management - bloat, concurrent
rebuilds, invalid states, and monitoring - rarely gets the spotlight it deserves.

This session dives into the maintenance work that keeps indexes healthy in long-running production systems. We'll cover how index bloat accumulates and why VACUUM alone won't save you, how to detect and measure bloat without expensive full-table scans, and the
right way to rebuild indexes using REINDEX CONCURRENTLY including the failure modes that can leave you with an invalid index at 2am. We'll also explore how HOT updates interact with index writes, why fill factor tuning is underused, and how to build a proactive index health monitoring practice using catalog queries that don't bring your database to its knees.
Attendees will leave with a practical maintenance playbook: query templates, alerting thresholds, and an operational checklist for keeping indexes lean and valid as your database grows.