Autotuning PostgreSQL on Azure Flexible Server
summary
Autotuning system built around DBtune open source agent connects to Azure Flexible Server via Azure CLI and Azure Monitor, iteratively changes pgconf, observes CPU/memory/latency and converges to better configs. On a TPC-C run on a D8ds baseline (≈550 TPS) the optimizer finds settings that reach ≈1,500 TPS — about a 2.5x throughput gain without resizing VMs. Production guardrails monitor memory and revert bad configs to keep stability.
AI-written summary · about
chapters (8)
- 0:00Intro
- 1:50What is database tuning?
- 5:25Database system parameter tuning
- 13:11How is parameter tuning tackled today
- 19:34Introducing DBTune
- 21:28DBTune Architecture on Azure
- 24:04Performance tuning results
- 25:58Safety in production environments
related talks
28mDear Azure Database for PostgreSQL, can you automate my index?
48mDBtune: AI-driven performance tuning for all PostgreSQL flavors
26mTop 10 reasons to choose Flexible Server for Azure Database for PostgreSQL
25mWhat’s New with Azure Database for PostgreSQL- Flexible Server
20mDebate: Autotuning
27mBuilding safety tooling for risk-free AI tuning of Postgres
27mBoosting Azure Database for PostgreSQL Performance with Azure Advisor
36mBest Practices in PostgreSQL Tuning: Navigating Key Performance Bottlenecks in the Cloud