Managing Kubernetes clusters is complex. Engineers spend countless hours running kubectl commands, analyzing logs, and debugging issues. What if an AI assistant could help you inspect your cluster, analyze problems, and provide insights—without any risk of accidental modifications?
The kubernetes-mcp project solves this exact challenge.
kubernetes-mcp is an open-source project that bridges Kubernetes clusters with AI assistants through the Model Context Protocol (MCP), ensuring 100% read-only safety.
Get started with kubernetes-mcp →
What is MCP?
Model Context Protocol (MCP) is an open protocol that enables AI assistants to safely interact with external systems through well-defined tools. Think of it as an API specifically designed for AI-human-system collaboration.
What is kubernetes-mcp?
kubernetes-mcp is a specialized MCP server built from the ground up for safe, read-only Kubernetes operations. Using Go and official Kubernetes client libraries, it gives AI assistants comprehensive visibility into your cluster while completely preventing any modifications.
Why Read-Only Matters
Traditional Kubernetes tools come with inherent risks when integrated with AI. A misinterpreted command could delete critical resources or modify production configurations. Kubernetes MCP eliminates these risks entirely:
- 🔒 Pure read-only architecture: No write operations exist in the codebase
- 🛡️ Production-safe by design: Impossible to accidentally modify resources
- ✅ Zero modification risk: Perfect for AI-assisted operations
- 🧠 AI-native integration: Purpose-built for intelligent assistants
Quick Start
Getting started with kubernetes-mcp takes just two steps:
# Install kubernetes-mcp
go install github.com/kkb0318/kubernetes-mcp@latest
Then configure your MCP-compatible AI assistant:
{
"mcpServers": {
"kubernetes": {
"command": "/path/to/kubernetes-mcp"
}
}
}
Four Powerful Tools for AI-Driven Insights
kubernetes-mcp provides four specialized tools that enable AI assistants to intelligently explore and analyze your cluster:
1. list_resources – Smart Resource Discovery
Discover resources across your entire cluster with intelligent filtering.
2. describe_resource – Deep Resource Analysis
Get detailed information about any resource, including configurations, status, and relationships. AI can analyze complex setups and identify misconfigurations.
3. get_pod_logs – Intelligent Log Analysis
Retrieve logs with time-based filtering, allowing AI to correlate events, identify patterns, and diagnose issues across multiple pods.
4. list_events – Event-Driven Debugging
Access cluster events to understand what happened and when. AI can reconstruct incident timelines and identify root causes.
Real-World Impact
The beauty of AI-powered exploration is its flexibility. While every AI model behaves differently, here’s what we’ve observed in practice:
Example: GitOps Health Check
You ask: „What GitOps resources are failing in my cluster?“
In our experience, the AI typically:
- Discovers all FluxCD/ArgoCD resources across namespaces
- Analyzes their current health status
- Identifies specific failures and their root causes
- Provides actionable remediation steps
What would normally require multiple kubectl commands and deep expertise happens instantly through natural language.
Example: Debugging a Failed Service
You ask: „Check the payment service status“
The AI might investigate by:
- Finding pods with „payment“ in their names
- Checking their current state
- Scanning logs for ERROR or FATAL messages
- Looking for related warning events
Based on the available information, the AI could identify issues like OOMKilled events, connection timeouts, or configuration errors – helping you quickly focus on the actual problems.
Who Benefits?
-
DevOps Engineers: Streamline debugging and reduce mean time to resolution
- „What pods are consuming the most memory?“
- „Show me all failing deployments in the last hour“
-
Developers: Understand deployments and troubleshoot issues without Kubernetes expertise
- „Why is my app not starting?“
- „What environment variables are set for my service?“
-
Platform Teams: Build AI-assisted operations tools with confidence
- „Generate a report of all deprecated API usage“
- „List all resources without resource limits“
-
On-Call Engineers: Get instant insights during incidents
- „What changed in the cluster in the last 30 minutes?“
- „Show me error patterns across all services“
Start Today
kubernetes-mcp is open source, production-ready, and designed for seamless integration with any MCP-compatible AI assistant. Experience the future of safe, intelligent Kubernetes operations.
Get started with kubernetes-mcp →
What’s Next?
kubernetes-mcp is under active development with new read-only tools being added regularly. We’re building a comprehensive suite of AI-powered Kubernetes inspection capabilities, driven by real-world needs and community feedback.
Join the Journey
- 🌟 Star the project to stay updated
- 💡 Open issues to suggest features
- 🤝 Contribute to the development
- 💬 Share your use cases