Prometheus MCP Server

Query Prometheus metrics and execute PromQL via standardized MCP interfaces.

Data last scanned today · Reviewed yesterday

456 tools★ 501PythonMIT

Overview

This server facilitates communication between AI assistants and Prometheus instances, enabling users to perform PromQL queries directly from their development environments. It supports integration with various MCP-compatible clients including Claude Desktop, VS Code, and Cursor. Installation is streamlined through Docker containers, allowing users to define their Prometheus URL as an environment variable within their client configurations. The server provides a bridge for AI to interpret and analyze metric data from your monitoring infrastructure.

Our verdict

With a security score of 30/100, this tool is classified as high-risk. A major concern is the total lack of authentication, meaning any process with access to the configuration can interact with your Prometheus instance. Additionally, there is no read-only mode to constrain the server's reach. Given these risks, this should only be utilized in isolated, low-stakes development environments where the Prometheus instance does not hold sensitive or production-critical data.

Zero direct dependencies reduce the supply-chain surface
Active repository with recent commits
Supported by seventeen contributors
Distributes via GHCR and Helm

No authentication methods implemented
Lacks support for OAuth
Community-maintained without official vendor support
Lacks read-only mode for query restriction

Tools

Tool	Description	Risk
health_check	Health check endpoint for container monitoring and status verification	low
execute_query	Execute a PromQL instant query against Prometheus	low
execute_range_query	Execute a PromQL range query with start time, end time, and step interval	low
list_metrics	List all available metrics in Prometheus with pagination and filtering support	low
get_metric_metadata	Get metadata for one metric or bulk metadata with optional filtering	low
get_targets	Get information about all scrape targets	low

Compatibility

Client	Local	Docker	Remote	Read-only
ChatGPT
Claude Desktop
Cursor
VS Code
Windsurf

Frequently asked questions

›What are the core prerequisites to use this server?

You need a Prometheus server that is accessible from your environment and an MCP-compatible client such as Claude Desktop, VS Code, Cursor, or Windsurf.

›How can I authenticate with my Prometheus instance?

The server supports basic authentication using the PROMETHEUS_USERNAME and PROMETHEUS_PASSWORD environment variables, as well as bearer tokens via PROMETHEUS_TOKEN. For mutual TLS, you can provide paths to your client certificate and private key using PROMETHEUS_CLIENT_CERT and PROMETHEUS_CLIENT_KEY.

›Is it possible to reduce the number of tokens used in AI responses?

Yes, you can set the environment variable PROMETHEUS_DISABLE_LINKS to True to prevent the inclusion of Prometheus UI links in query results.

›How do I prevent the server from hanging on slow queries?

You can configure the PROMETHEUS_REQUEST_TIMEOUT environment variable to specify a timeout in seconds, which defaults to 30 seconds.

›Can I disable SSL verification for my Prometheus connection?

Yes, you can disable SSL verification by setting the PROMETHEUS_URL_SSL_VERIFY environment variable to False.