All experiments
AgentsKubernetesDevOps
Kubernetes AI Assistant
Natural-language assistant that inspects cluster state, suggests fixes, and generates safe kubectl / Helm workflows.
Problem
Cluster troubleshooting is context-heavy. Operators jump between dashboards, logs, and docs while under incident pressure.
Architecture
- 1Chat UI with tool-call timeline
- 2AI gateway with policy filters
- 3Read-only Kubernetes API tools
- 4Optional write tools behind confirmation
Screenshots
Screenshot placeholder 1
Screenshot placeholder 2
Lessons Learned
- Read-only tools should be the default; writes need human confirmation.
- Structured tool schemas beat free-form shell suggestions for reliability.
- Grounding answers in live cluster facts reduces hallucinated commands.
Technologies
GoKubernetesOpenAILangChain