Bloat

Bloat is the accumulation of dead space within a PostgreSQL table or index, caused by dead tuples not yet removed by VACUUM. A table with 50% bloat occupies twice the necessary space and forces sequential scans to read twice as many pages.

How it works #

Bloat is measured by comparing the actual table size with the expected size based on live rows. The pgstattuple extension provides the dead_tuple_percent field. Bloat above 20-30% is a warning sign; above 50% is an emergency.

What it’s for #

Monitoring bloat is essential to understand whether autovacuum is keeping pace. The pg_stat_user_tables query with n_dead_tup and last_autovacuum is the first diagnostic tool. If bloat is out of control, pg_repack rebuilds the table online without prolonged exclusive locks — unlike VACUUM FULL.

What can go wrong #

Normal VACUUM reclaims dead tuple space but doesn’t compact the table — fragmented space remains. If bloat reaches 50-70%, VACUUM alone isn’t enough. The options are VACUUM FULL (exclusive lock, blocks everything) or pg_repack (online, but requires installation). The real solution is not getting there, with a well-configured autovacuum.