datadog

This MCP server exposes Datadog resources through the MCP (Model Context Protocol) interface. It provides tools to interact with Datadog's monitoring data, dashboards, metrics, events, logs, and incidents via MCP calls such as get-monitors, get-dashboard, search-logs, and aggregate-logs. You can use the available tools to fetch, filter, and aggregate data from your Datadog account, enabling integrations with other MCP-enabled tools or assistants. The server supports service-specific endpoints for logs and metrics, and it includes error handling for authentication and API issues to help you diagnose problems quickly.

Prerequisites:

• Node.js version 16 or higher installed on your system
• npm (comes with Node.js) or pnpm/yarn if you prefer

Installation steps:

1. Install the MCP server globally via npm (recommended):

npm install -g datadog-mcp-server

2. Ensure you have a Datadog account with an API key and an Application key. These keys can be created in Datadog under Organization Settings > API Keys and Application Keys.

3. Start the server with environment variables (recommended) or pass keys as CLI arguments:

# Using environment variables
export DD_API_KEY=your_api_key_here
export DD_APP_KEY=your_app_key_here
export DD_SITE=datadoghq.com

# Start the MCP server (npx form shown for completeness)
npx datadog-mcp-server

4. Optional: run via Claude Desktop config as shown in the README for quick CLI integration.

Notes and tips:

• You can scope your Datadog Application Key to only the permissions needed by the MCP server to follow the principle of least privilege.
• If you operate across regions, you may need to set DD_SITE (and optionally DD_LOGS_SITE, DD_METRICS_SITE) to EU or GovCloud endpoints as appropriate.
• The server exposes a suite of MCP tools including get-monitors, get-dashboards, get-metrics, get-events, search-logs, aggregate-logs, and get-incidents. Use the examples in the README as a reference for the exact tool call structure.
• In Claude Desktop, you can configure the mcpServers section to run the server via npx datadog-mcp-server with the necessary apiKey/appKey and site settings.
• If you encounter 403 errors, verify that keys have the required scopes (monitors_read, dashboards_read, metrics_read, events_read, logs_read_data, incident_read) and that you are using the correct regional endpoint.
• The server supports both environment-variable-based configuration and command-line arguments for flexibility in deployment environments.