Data-intensive PostgreSQL: Three ways to scale
summary
Explains three concrete scaling patterns with tools and tradeoffs: for multi‑tenant SaaS use schema-based sharding or row‑per‑tenant to keep per‑tenant ops local and isolate large customers, and prefer COPY over single‑row inserts for high ingest rates; for real‑time dashboards use partitioned time‑series with pre‑aggregation and consider Citus to parallelize insert‑selects so minute‑latency dashboards stay fast; for BI use columnar storage plus vectorized execution (Citus columnar, pg_analytics/DataFusion, DuckDB, Crunchy Bridge) with object storage for massive volumes.
AI-written summary · about
chapters (6)
- 0:00Intro
- 2:40How to scale Postgres
- 5:51Multi-tenant SaaS applications
- 11:35User-facing time series dashboards
- 16:38Business intelligence
- 22:37Summary
related talks
33mFive Sharding Data Models and Which is Right
37mPostgres at Any Scale
36mPostgres at any Scale
42mPostgreSQL Distributed: Architectures & Best practices
51mBuilding a scalable time-series database on PostgreSQL
51mPostgreSQL, The Time-Series Database You Want
30mManaging Postgres at scale: Challenges, Tools & Techniques
28mDivide and Conquer: Multi-tenancy in Postgres