Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a prototype system that enables Presto to automatically diagnose and fix failed queries using AI-powered analysis. Learn how this innovative approach captures failed queries caused by memory limits, skewed joins, or connector issues, then leverages Large Language Models (LLMs) to analyze failure patterns and automatically suggest or retry mitigated versions. Discover the technical implementation including query log analysis, EXPLAIN plan utilization, retry logic mechanisms, and Presto session properties configuration. Understand how the system can automatically add session configurations, break complex queries into smaller components, or rewrite problematic joins to resolve common failure scenarios. Gain insights into this forward-thinking approach that transforms Presto from a fast query engine into an intelligent, self-healing system capable of autonomous troubleshooting and recovery, reducing the manual intervention typically required when queries fail in production environments.
