Agentkube is an AI-powered Kubernetes management environment designed to simplify cluster operations and troubleshooting. It combines graphical resource management with intelligent agents that investigate infrastructure problems and suggest solutions. Its multi-agent system analyzes cluster states, logs, events, and dependencies to identify potential root causes. Users can connect more than 20 AI providers or run local models through Ollama and vLLM. A human-in-the-loop approval system provides control over actions proposed by agents. Integrated LogQL and PromQL query builders support log exploration and metrics analysis. Custom dashboards visualize workload health and cluster performance. Additional tools include container vulnerability scanning, resource recommendations, dependency maps, and MCP integrations for extending operational workflows.
Features
- AI-driven Kubernetes troubleshooting with multi-agent root cause analysis and automated issue investigation
- Support for more than 20 cloud AI providers and local models through Ollama and vLLM
- Human approval controls for agent actions with optional read-only cluster investigation mode
- Integrated LogQL and PromQL query builders for analyzing logs, events, and Prometheus metrics
- Interactive Kubernetes resource maps, customizable monitoring dashboards, and container vulnerability scanning
- Built-in resource optimization recommendations with Grype, Trivy, ArgoCD, and MCP marketplace integrations