Multi Cluster Kubernetes MCP Server

What is Multi Cluster Kubernetes MCP Server?

An MCP (Model Context Protocol) server for managing multiple Kubernetes clusters simultaneously. It provides 60+ tools covering cluster operations, resource management, monitoring, RBAC, storage, networking, and more, accessible through AI assistants like Claude Desktop. Designed for DevOps engineers and platform teams managing multiple Kubernetes clusters.

How to use Multi Cluster Kubernetes MCP Server?

Install via Smithery with npx -y @smithery/cli install @razvanmacovei/k8s-multicluster-mcp --client claude, or clone the repository and run python3 app.py. Configure the KUBECONFIG_DIR environment variable to point to a directory with kubeconfig files; the server automatically discovers all contexts. Required: Python 3.11+, valid kubeconfig files, and metrics-server for monitoring tools.

Key features of Multi Cluster Kubernetes MCP Server

• Automatic discovery of all Kubernetes contexts across kubeconfig files.
• Partial name matching for cluster selection (e.g., prod).
• 60+ tools for cluster health, resource CRUD, scaling, rollouts, RBAC, storage, networking.
• Cross-cluster operations: compare resources, health, and configurations.
• Context caching with 30-second TTL for performance.

Use cases of Multi Cluster Kubernetes MCP Server

• Get a health overview of a production cluster via an AI assistant.
• Compare pod counts in the backend namespace between prod and staging.
• Diagnose application issues in a deployment with automated recommendations.
• Scale a deployment to a specific number of replicas.
• Manage secrets, ConfigMaps, and RBAC across multiple clusters.

FAQ from Multi Cluster Kubernetes MCP Server

What is the difference between this server and using kubectl directly?

This server provides an MCP interface for AI assistants to manage multiple clusters simultaneously, with automatic context discovery and cross-cluster comparison capabilities, unlike direct kubectl usage.

What are the runtime requirements?

Python 3.11 or higher, one or more kubeconfig files with valid cluster credentials, and metrics-server installed in clusters for k8s_top_* and k8s_cluster_info tools.

Where are kubeconfig files located?

By default, the server reads from ~/.kube. You can set the KUBECONFIG_DIR environment variable to any directory containing kubeconfig files.

Are there any known limitations?

Namespace deletion protects system namespaces. The server requires metrics-server for monitoring tools. Pod exec supports quoted arguments fixed in v2.0.0.

What transport or authentication does it use?

The server uses the MCP protocol over stdio (local execution). Authentication relies on the Kubernetes credentials provided in the kubeconfig files.