PostgreSQL performance tips you have never seen before
summary
Shows concrete wins: switching from TCP localhost to a UNIX socket doubles a trivial SELECT 1 throughput (100,000→200,000 tps), packing fixed-size columns before varchars shrinks a sample table (651MB→574MB), setting defaults smartly reduces row storage (420MB→346MB), replacing a text index with a hashed functional index cuts an email index from 825MB to roughly a quarter of that, tuning FDW fetch_size drops 10M-row reads (5.2s→3.6s), and frequent VACUUM for GIN indexes turns 9ms queries into sub-millisecond responses.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 2:25Connecting to Postgres
- 6:25Storing data efficiently
- 10:48Considering indexes
- 14:10Composite type trickery
- 17:07Using Foreign Data Wrappers
- 19:49Full Text Indexes
related talks
26mPostgreSQL performance tips you have never seen before V2.0
42mPostgreSQL Tips & Tricks For App Devs
52mBoosting Typical Query Patterns: PostgreSQL 18's Performance Enhancements
48mModern SQL features in PostgreSQL
25mScaling PostgreSQL Beyond Query Optimization
47mPostgreSQL worst practices
50mImproving Postgres' Efficiency
56mPerformance Improvements in PostgreSQL9 2