Datadog MCP Server

April 14, 2026 · View on GitHub

Datadog's managed MCP server connects your AI agents to Datadog's observability tools and data. Query logs, metrics, traces, incidents, and more - directly from your AI-powered dev tools.

Quick Setup

For Claude Code (US1 site):

claude mcp add --transport http datadog https://mcp.datadoghq.com/api/unstable/mcp-server/mcp

Many MCP clients also support configuration in a .mcp.json file:

{
  "mcpServers": {
    "datadog": {
      "type": "http",
      "url": "https://mcp.datadoghq.com/api/unstable/mcp-server/mcp"
    }
  }
}

Not on the US1 site? Replace mcp.datadoghq.com with the endpoint for your Datadog site — for example, EU users should use mcp.datadoghq.eu.

The MCP server should work with any standards-compliant MCP client. See the full setup docs for other clients, alternative auth methods, and regional endpoints.

Automated Agent Examples

examples/claude-agent-sdk/ — Minimal agent using the Claude Agent SDK
examples/claude-agent-sdk-advanced/ — Monitoring agent that investigates issues and posts to Slack. Includes Datadog LLM Observability integration

Quick Setup

Automated Agent Examples

Links