The self-driving database: Automating Postgres tuning with Bayesian optimization

Database tuning is often a manual bottleneck for platform teams. This talk introduces an automated framework that uses Bayesian optimization (TPE) to auto-tune PostgreSQL configurations. The speaker breaks down a two-stage optimization pipeline that satisfies strict recovery time objectives (RTO) while autonomously exploring parameter spaces to maximize TPS. Platform engineers gain a blueprint for implementing "self-driving" data infrastructure that scales performance while reducing operational toil.