MCP Server Kubernetes

Name: MCP Server Kubernetes
Author: Flux159

@Flux159

リポジトリホームページ

MCP Server Kubernetes について

MCP Server for kubernetes management commands

#kubernetes #infrastructure #server #mcp

基本情報

カテゴリ

クラウドとインフラ

ライセンス

MIT

ランタイム

node

トランスポート

stdio

公開者

Flux159

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "kubernetes": {
      "command": "npx",
      "args": [
        "mcp-server-kubernetes"
      ]
    }
  }
}

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is MCP Server Kubernetes?

MCP Server Kubernetes is a Model Context Protocol server that connects to a Kubernetes cluster and enables AI assistants to manage it. It loads kubeconfig from multiple sources in priority order and integrates with clients such as Claude Code, Claude Desktop, VS Code, Cursor, and Gemini CLI. It is intended for developers who need AI-assisted Kubernetes operations.

How to use MCP Server Kubernetes?

Install it via npx (e.g., npx mcp-server-kubernetes) and configure the server in your MCP client’s settings (e.g., claude mcp add kubernetes -- npx mcp-server-kubernetes). The server automatically uses your current kubectl context. Prerequisites: kubectl in PATH, a valid kubeconfig, and cluster access; Helm v3 is optional.

Key features of MCP Server Kubernetes

Connects to any Kubernetes cluster accessible via kubectl.
Unified kubectl API: get, describe, create, apply, delete, logs, context, explain, scale, patch, rollout, and generic commands.
Advanced operations: port forward, Helm install/upgrade/uninstall, pod cleanup, node management.
Non-destructive mode that blocks destructive operations.
Secrets masking for security.
Optional OpenTelemetry observability with distributed tracing.

Use cases of MCP Server Kubernetes

Manage Kubernetes resources through natural-language chat prompts in AI assistants.
Automatically diagnose pod issues using the built-in /k8s-diagnose troubleshooting flow.
Safely monitor and inspect clusters in read‑only or create/update‑only mode.
Perform Helm chart operations (install, upgrade, uninstall) via AI-guided commands.

FAQ from MCP Server Kubernetes

What are the prerequisites for running MCP Server Kubernetes?

You need kubectl installed and in your PATH, a valid kubeconfig file with configured contexts, and access to a Kubernetes cluster (e.g., minikube, Rancher Desktop, GKE). Helm v3 is optional.

How do I enable non-destructive mode?

Set the environment variable ALLOW_ONLY_NON_DESTRUCTIVE_TOOLS=true when starting the server. This disables all destructive operations (delete, drain, etc.) while keeping read and create/update tools available.

Does MCP Server Kubernetes support observability?

Yes, it includes optional OpenTelemetry integration (disabled by default). Enable it with ENABLE_TELEMETRY=true and configure OTEL_EXPORTER_OTLP_ENDPOINT. Traces can be sent to Jaeger, Grafana Tempo, Datadog, New Relic, and other OTLP-compatible backends.

Where does the server load its configuration from?

By default, it loads kubeconfig from ~/.kube/config. Additional authentication options (environment variables, custom paths) are described in the advanced documentation.

What transport and authentication does the server use?

It uses stdio transport configured through your MCP client. Authentication relies on your existing kubectl context and kubeconfig; no separate credentials are managed by the server itself.

「クラウドとインフラ」の他のコンテンツ

AWS Model Context Protocol (MCP) Server

alexei-led

A lightweight service that enables AI assistants to execute AWS CLI commands (in safe containerized environment) through the Model Context Protocol (MCP). Bridges Claude, Cursor, and other MCP-aware AI tools with AWS CLI for enhanced cloud infrastructure management.

184