The Seven Principles for Making a PostgreSQL DB Work Fast
summary
Seven practical rules: ground schema in the real business model to avoid leaky abstractions; run PG stat statements and auto_explain to surface slow queries and capture plans; balance work between stored procedures and app code based on latency, team SQL fluency and hosting limits; design multicolumn indexes only after understanding query patterns; scale up, then add replication or sharding (Citus-style) when needed; use materialized views or ETL to columnar DBs for heavy analytics; prefer managed Postgres where it reduces ops.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:13Principle 1 Good Relations
- 1:30Principle 2 Profiling
- 2:51Principle 3 Data Processing
- 4:52Principle 4 Indexing
- 6:29Principle 5 Scaling
- 7:53Principle 6 Quitting Relation Game
- 9:09Principle 7 Using the Right Tools
- 10:02Principle 8 Know Your Team
related talks
46mPostgres Platform "Best Practices" for the Modern DBA
41mOptimizing your app by understanding your PostgreSQL database
47mPostgreSQL worst practices
37mPostgres at Any Scale
31mPostgres Performance for Humans
42mPostgreSQL Tips & Tricks For App Devs
57mPostgreSQL When It's Not Your Job
49mBeyond PostgreSQL 17