Agentic Incident Response
summary
An agent runs a reasoning loop over observability data (Elastic, InfluxDB, Grafana) and repo history via MCP, finds a memory leak introduced by a recent cache PR, and executes mitigations: it patches Kubernetes memory limits, updates the status page, and opens a pull request that fixes three bugs including missing cache invalidation. The stack is model-agnostic with an LLM orchestration layer using GPT5 and Cloud Standard 4.5, schema-driven reasoning, and about 20 scoped tools per loop.
AI-written summary · about
chapters (8)
- 0:00Introduction and motivation
- 3:05Current incident response challenges
- 8:49Live demo of agentic root cause analysis
- 18:13Why now for agentic incident response
- 20:17Building an AI SRE agent
- 23:50Lessons learned with MCP servers
- 25:59Agent architecture and future plans
- 27:00Q&A session
related talks
35mAgentic Incident Response
30mAI Agents for Incident Management
40mWhen Incidents Fix Themselves: AI SRE in action
36mModernizing Incident Response with LLMs, RAG, and the MCP
15mIntelligent Incident Management in DevOps
16mAI Agents for Incident Investigation: The Good, The Bad, and The Ugly
27mAccelerating Incident Response With OpenTelemetry and MCP: Building...
28mAI-Driven Incident Response: From Alert to Postmortem