Multi-Agent AI System Investigating Kubernetes Incidents Automatically(opsworker.ai)
opsworker.ai
Multi-Agent AI System Investigating Kubernetes Incidents Automatically
https://www.opsworker.ai/blog/agent-driven-sre-investigations-a-practical-deep-dive-into-multi-agent-incident-response/
2 comments
I’ve been exploring how far we can push fully autonomous, multi-agent investigations in real SRE environments — not as a theoretical exercise, but using actual Kubernetes clusters and real tooling. Each agent in this experiment operated inside a sandboxed environment with access to Kubernetes MCP for live cluster inspection and GitHub MCP to analyze code changes and even create remediation pull requests.