// the find
ankane/pghero
A performance dashboard for Postgres
PgHero is a Postgres performance dashboard, shippable as a standalone Docker image or mounted into a Rails app as an engine. It surfaces slow queries, index suggestions, space/bloat, connections, and replication lag in one place. Good fit for teams running Postgres who want an at-a-glance operational view without standing up a full observability stack.
Covers a wide range of operational concerns in one tool — slow queries via pg_stat_statements, index suggestions, table/index bloat, connection counts, replication lag, autovacuum status — so you're not stitching together five separate scripts. Works read-only against Postgres system catalogs and stats views, so it's low-risk to point at production; nothing runs as an agent on the DB server itself. Ships two deployment paths (Docker image or Rails engine), so it fits both Rails and non-Rails shops. Maintained by ankane with a long track record in this space (same author as Dexter and pgsync) and real production use at Instacart, which is a reasonable signal of reliability.
The most useful feature, query stats, depends on pg_stat_statements being enabled with the right track settings and preload config — plenty of managed Postgres tiers restrict this, so the headline feature may just not work depending on where you're hosted. Suggested indexes are heuristic, not based on actual query plans the way Dexter's are, so recommendations can miss real workload patterns or suggest indexes that don't help. No built-in authentication — you're expected to wrap it in your own Rails auth or stick it behind a reverse proxy, which is easy to forget if you deploy the standalone Docker image directly. The UI is functional but visibly dated (Bootswatch/Chart.js-era look), so don't expect it to double as something you'd show off.