Best Practices in PostgreSQL Tuning: Navigating Key Performance Bottlenecks in the Cloud
summary
Cloud deployments show big, measurable effects: a cross‑region pgbench run yields ~3,000 transactions vs ~500,000 in same‑zone with avg latency ~1s vs ~23ms, and TPS drops by two orders of magnitude, so colocate app and DB. Flexible Server adds PG bouncer on the VM, SSD V2 and V5 compute give much higher throughput; a third‑party benchmark reports ~4.7× ops/min, ~2.86× IOPS and ~3.8× performance‑per‑dollar over single server. Limit client connections (start at CPUs × 5), use pgbench to validate, and migrate to flexible server.
AI-written summary · about
chapters (8)
- 0:00Introduction to Azure Postgres
- 2:03Geography and latency
- 11:22Benchmarking results
- 13:48Hardware and storage choices
- 16:21Software and migration
- 19:52AMD performance analysis
- 24:52Migration and connectivity
- 26:40Managing connections
related talks
43mPostgreSQL in the cloud: Theory and Practice
26mTop 10 reasons to choose Flexible Server for Azure Database for PostgreSQL
36mAutomated Performance Testing of Postgres
28mRun PostgreSQL in Warp Speed Using NVMe/TCP in the Cloud
46mUnlock PostgreSQL’s True Performance with Local SSDs
48mModern SQL features in PostgreSQL
46mBig databases, many servers, data center or cloud - it's all here!
50mPerformance Archaeology