You can connect Claude Desktop App to your Kubernetes cluster through Metoro MCP Server to ask questions and receive context-rich insights from your cluster telemetry. This server exposes APIs that let your AI assistant access Metoro’s telemetry data for interactive exploration, troubleshooting, and observations without manual querying.

How to use

You interact with Metoro MCP Server through an MCP client like Claude Desktop App. In Claude, select Metoro MCP Server as your MCP target. Provide the required authentication and API URL so the client can fetch cluster telemetry and respond to your questions. You can ask about pod health, deployment status, resource usage, and custom metrics collected by Metoro, and the AI will use that data to generate actionable answers.

How to install

Prerequisites you need before starting are: the Claude Desktop App and the Go toolchain.

Install Go on your system if it is not already installed.

Clone the MCP server repository, build the binary, and prepare to run it.

Step by step commands

# Prerequisites
# Install Claude Desktop App from your preferred distribution
# Install Go (if not already installed)
# macOS: brew install go
# Linux: sudo apt-get install -y golang

# Clone the MCP server repository
git clone https://github.com/metoro-io/metoro-mcp-server.git

# Build the server executable
cd metoro-mcp-server
go build -o metoro-mcp-server

# If you have a Metoro account with an auth token:
# Create the Claude config file with your token and API URL

{ "mcpServers": { "metoro-mcp-server": { "command": "<your path to Metoro MCP server go executable>/metoro-mcp-server", "args": [], "env": { "METORO_AUTH_TOKEN" : "<your auth token>", "METORO_API_URL": "https://us-east.metoro.io" } } } }

If you want to try the Live Demo Cluster without a Metoro account:

Use the provided public demo token and demo API URL

{
  "mcpServers": {
    "metoro-mcp-server": {
      "command": "<your path to Metoro MCP server go executable>/metoro-mcp-server",
      "args": [],
      "env": {
          "METORO_AUTH_TOKEN" : "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJjdXN0b21lcklkIjoiOThlZDU1M2QtYzY4ZC00MDRhLWFhZjItNDM2ODllNWJiMGUzIiwiZW1haWwiOiJ0ZXN0QGNocmlzYmF0dGFyYmVlLmNvbSIsImV4cCI6MTgyMTI0NzIzN30.7G6alDpcZh_OThYj293Jce5rjeOBqAhOlANR_Fl5auw",
          "METORO_API_URL": "https://demo.us-east.metoro.io"
       }
    }
  }
}

Connect the client and start using

Save the configuration file, then restart the Claude Desktop App. You should see Metoro MCP Server appear in the MCP Server dropdown. Select it, ensure the token and API URL are correct, and begin asking questions about your Kubernetes cluster.

Notes and best practices

• Keep your METORO_AUTH_TOKEN secure. Treat it as a sensitive credential and rotate it as needed.
• The API URL points to Metoro’s backend you’re querying against; ensure it is reachable from your environment.
• You can switch between a real Metoro account and the demo cluster depending on your testing needs.

Security and access

Only authenticated requests with a valid token are allowed. Do not expose your auth token in shared files or public spaces.