PostgreSQL: Tuning parameters or Tuning Queries? with Henrietta Dombrovskaya
summary
Claims that tuning Postgres parameters rarely beats fixing queries and indexes, demonstrated on the posgressier dataset from GitHub using actual timings: a sample query runs 2.4s on a clean DB, parameter tweaks (max_parallel_workers_per_gather, work_mem, shared_buffers) shave it toward ~1.1s, but creating two appropriate indexes drops it to 0.06s — roughly a 20x speedup — and removing the parameter tweaks changes nothing. Parameters exist to tell Postgres about cores and memory; indexes and query changes deliver the large wins.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 2:31Talk about tuning parameters
- 6:49Query vs parameter tuning
- 12:32Real purpose of parameters
- 15:02Q&A
- 17:10App performance recommendations
related talks
18mTuning Parameters in Postgres vs. Tuning Your Queries
43mPostgres Performance: From Slow to Pro
04mPostgres Tuning is saved by Sovenath Shaw
45mProper PostgreSQL Parameters to Prevent Poor Performance
23mBest Practices for Tuning Slow Postgres Queries
24mTop 5 PostgreSQL Query Tuning Tips
27mPostgres Performance in 15 Minutes
34mTuning PostgreSQL to work even better