Sponsored Keynote: Scaling Platform Ops with AI Agents: Troubleshootin... Jorge Palma & Natan Yellin

This keynote covers the deployment of AI agents for maintaining Kubernetes clusters at scale, presented by Jorge Palma from Microsoft and Natan Yellin from Robusta. The speakers introduce HolmesGPT, an open-source CNCF sandbox project that utilizes large language models to analyze operational and observability data for diagnosing production issues. The talk addresses the transition from diagnosis to remediation, illustrating how agents can autonomously detect and resolve issues while adhering to strict Role-Based Access Control, approval workflows, and audit trails. The speakers emphasize that this technology is designed to enhance the effectiveness of Site Reliability Engineers, enabling them to concentrate on creative problem-solving and system design rather than routine troubleshooting.