AI Agents for Incident Management
summary
An AI responder agent that runs a reasoning loop over observability and code is built and demoed: it integrates via the model context protocol with sources like Grafana and Loki, inspects deployments, logs and repo commits, and claims to cut root cause analysis from ~30 minutes to ~1 minute. Live demos show it auto-detecting a SQL typo and generating a fix and status update (works), while a staged Kubernetes memory leak exposes limits, occasional hallucination and missing tool access (fails until tools and permissions are enabled).
AI-written summary · about
chapters (7)
- 0:00Welcome and Introduction
- 0:38The Role of AI in Incident Management
- 1:42Current Challenges and Future Vision
- 2:05Building an AI Agent for SREs
- 2:28Interactive Q&A Session
- 15:12AI in Action: Live Demonstration
- 27:04Conclusion and Final Thoughts
related talks
35mAgentic Incident Response
16mAI Agents for Incident Investigation: The Good, The Bad, and The Ugly
34mAgentic Incident Response
40mWhen Incidents Fix Themselves: AI SRE in action
19mAI Agents for Incidents: The Good, the Bad, and the Ugly
17mFighting AI with AI
42mAI for security - friend or foe?
30mAI and Incident Management: How to Reduce Manual Work and Speed up the Incident...