Root Cause Analysis Using Large Language Models - Beyond Vitrage

Explore a 24-minute conference talk that introduces an innovative approach to Root Cause Analysis (RCA) using Large Language Models (LLMs) to overcome the limitations of OpenStack's Vitrage service. Learn how this enhanced solution maps relationships between OpenStack resources and services, leveraging tracing tools like Jaeger to monitor service interactions and identify bottlenecks. Discover how the integration of LLMs through fine-tuned prompts and customized error scenarios significantly improves the RCA process. Understand the expanded capabilities that come from integrating with Kubernetes, Prometheus, and other monitoring tools, creating a more comprehensive and scalable approach to RCA in complex cloud environments. Presented by Chaeri Kim and Wooyoung Choi, gain insights into this groundbreaking advancement in cloud infrastructure management that promises more efficient problem diagnosis and resolution.