Smarter Databases in Kubernetes: Applying AI for Query Optimization and Self-Tuning
summary
Argues that Kubernetes is a natural control plane for adaptive databases and that current systems sit around level two autonomy, not fully autonomous, with a surprising poll result of only 5% trust AI to run DB completely. Describes practical tooling and experiments: Releem and PGAnalyze for index and config tuning, EverSQL/Aiven for query rewrites, HyperPG for safe index trials, and Open Everest (CNCF sandbox) as an operator-backed RDS-like control plane. Emphasizes policies for safety and a mix of control-plane and future in-kernel AI.
AI-written summary · about
chapters (12)
- 0:00AI-driven databases in Kubernetes
- 3:34Kubernetes for smart databases
- 8:09AI in database operations
- 10:37Core database problems AI solves
- 13:23AI in database optimization
- 16:14AI-driven configuration tuning
- 17:59Schema and index optimization with AI
- 20:05Scaling databases and costs
- 21:51Requirements for AI in databases
- 23:18AI autonomy vs human oversight
- 24:44Future of self-driving databases
- 25:36Q&A
related talks
06mHow to Make Kubernetes Smarter
25mExtending Kubernetes for AI | Lessons Learned From Platform Engineering
26mChallenges of Managing Stateful Applications in Kubernetes
11mDatabase Evolution: Partitioning and Indexing
25mPostgres on Kubernetes for the Reluctant DBA
29mDatabases with Kubernetes
07mLightning Talk: My Database Runs on Kubernetes. What's Next? Data Platforms!
38mBeyond Kubernetes: Adapting To Spe...