pab1it0/prometheus-mcp-server

What it solves

It allows AI assistants to interact directly with Prometheus metrics. Instead of a human manually running queries in a dashboard, an AI can execute PromQL queries and analyze metrics data to help with monitoring and troubleshooting.

How it works

The project implements a Model Context Protocol (MCP) server. It acts as a bridge between an MCP-compatible AI client (like Claude Desktop, Cursor, or VS Code) and a Prometheus server. The server exposes a set of standardized tools that the AI can call to perform actions such as executing instant or range queries, listing available metrics, and retrieving metric metadata.

Who it’s for

DevOps engineers, SREs, and developers who use Prometheus for monitoring and want to enable their AI coding assistants or LLM-based tools to query and analyze their system metrics.

Highlights

• PromQL Integration: Supports both instant and range queries for time-series data.
• Metric Discovery: Tools to list metrics and fetch metadata, allowing the AI to explore the available data without prior knowledge.
• Flexible Deployment: Can be deployed via Docker, Helm charts for Kubernetes, or as a standalone Python application.
• Enterprise-Ready Auth: Supports Basic authentication, Bearer tokens, and mutual TLS (mTLS) for secure connections to Prometheus.
• Configurable Toolset: Allows users to toggle which tools are exposed to the AI to optimize the context window.