Metoro MCP Server is an MCP server designed to connect Claude Desktop App with Kubernetes clusters, enabling AI-driven monitoring and management. It leverages Metoro's API to provide LLMs with real-time Kubernetes cluster data, offering insights into your infrastructure through natural language queries. By utilizing eBPF-based microservice observation, it generates deep telemetry data without requiring code modifications.

This server seamlessly integrates with AI models via the MCP protocol. To integrate, configure the claude_desktop_config.json file with the server's command, arguments, and environment variables, including your Metoro auth token. The core value lies in simplifying Kubernetes cluster management by bridging the gap between observability data and AI-powered analysis, allowing developers to quickly diagnose issues and optimize performance using natural language. Built with Golang and the Golang MCP SDK, it exemplifies a robust and efficient MCP server implementation.

Kubernetes Observability via LLM

Metoro MCP Server bridges the gap between Kubernetes cluster monitoring and Large Language Models (LLMs) by providing a natural language interface to query cluster state. It leverages Metoro's API to furnish LLMs with real-time Kubernetes observability data, enabling users to ask questions about their cluster's health, performance, and configuration using natural language. This eliminates the need for complex command-line queries or navigating dashboards, making Kubernetes management more accessible to a wider range of users. For example, a user could ask "What is the CPU usage of my web application?" and receive a concise, human-readable answer derived from the cluster's metrics. This functionality is achieved by translating the natural language query into a Metoro API call, retrieving the relevant data, and presenting it to the user in an understandable format.

eBPF-Powered Microservice Insights

A key feature of Metoro MCP Server is its ability to provide deep telemetry data for microservices without requiring any code modifications. This is achieved through the use of eBPF (Extended Berkeley Packet Filter) technology, which allows for the dynamic instrumentation of applications at runtime. By leveraging eBPF, Metoro MCP Server can automatically collect detailed performance metrics, trace requests, and identify bottlenecks within microservice architectures. This data is then made available to the LLM, providing a comprehensive view of the microservice's behavior. For instance, a developer could ask "What are the slowest database queries in my payment service?" and receive a list of queries along with their execution times, enabling them to quickly identify and address performance issues. This feature significantly reduces the overhead associated with traditional monitoring approaches, which often require manual instrumentation and code changes.

Seamless Integration with Claude

Metoro MCP Server is designed for seamless integration with the Claude Desktop App, a popular LLM-powered assistant. The integration is achieved through the Model Context Protocol (MCP), a standardized protocol for connecting AI models with external data sources and services. By adhering to the MCP standard, Metoro MCP Server can easily be configured to provide Claude with access to Kubernetes observability data. The configuration involves specifying the command, arguments, and environment variables required to launch the Metoro MCP Server within the claude_desktop_config.json file. This allows users to interact with their Kubernetes clusters using natural language directly within the Claude Desktop App, streamlining their workflow and enhancing their productivity. The server utilizes the Golang MCP SDK to facilitate communication and data exchange with the Claude client.