Postgres 10, Performance, and You
summary
Postgres 10 brings concrete wins: native table partitioning that lets bulk loads and deletes avoid vacuum bloat, hash indexes becoming crash‑safe and replicated, and native full‑text search over JSON. Parallel query work adds parallel index scans, gather_merge to keep worker context, and parallel merge join, which turns a 0.5GB repo table from a slow sequential scan into an index-backed plan with huge latency drops. Heroku pairs this with managed credentials, PGX hardware tiers, PG diagnose, and a PgBouncer beta for pooling.
AI-written summary · about
chapters (29)
- 0:00Intro
- 0:42Overview
- 2:45Inspirational Quotes
- 8:28Pivot
- 13:34JSON Blob
- 13:49Dynamic Row Sizes
- 14:49Full Text Index Searching
- 15:26Views and Perspectives
- 16:06Quotes
- 17:29Parallel Searching
- 18:40Gatherers
- 20:45Worker Collection Results
- 21:24Revisiting Worker Collection
- 21:35Keeping Context
- 22:04Mental Map
- 22:29Merge Joint
- 23:22Parallel Merge Joint
- 24:27RealWorld Example
- 25:03Query Analyzer
- 26:19Postgres 10 with Rails
- 27:13Heroku Postgres
- 27:52Postgres Extensions
- 28:16Postgres Manage credentials
- 28:47Customizing Manage credentials
- 29:12External control access
- 29:41Roku PGX plans
- 30:32Heroku PG diagnose
- 31:21Heroku PG bouncer
- 32:05Heroku stats
