mcp-datadog

This MCP server exposes tools for interacting with Datadog resources (Logs, Monitors, Metrics, Incidents, and Dashboards) via the MCP protocol. It supports a range of Datadog operations such as searching logs, listing and retrieving monitors, querying metrics, and inspecting incidents and dashboards. After starting the server (preferably using Docker as recommended), you can connect an MCP-enabled client (e.g., Claude Desktop/Code or another MCP client) and issue natural language or structured queries to the available tools. The server translates those requests into Datadog API calls and returns structured results suitable for downstream AI agents to reason about.

Prerequisites:

• Docker and Docker Compose installed on your machine
• Optional: .NET SDK if you prefer running the server directly (not required with the Docker approach)
• Datadog API Key and Datadog Application Key with appropriate read permissions for Logs, Monitors, Metrics, Incidents, and Dashboards

Step-by-step installation:

1. Clone the repository and navigate to the project:

git clone https://github.com/viamus/mcp-datadog.git
cd mcp-datadog

2. Create and configure environment file (copy example and edit):

cp .env.example .env

Edit .env to include your keys:

DATADOG_API_KEY=your-api-key-here
DATADOG_APPLICATION_KEY=your-application-key-here
DATADOG_SITE=api.datadoghq.com

3. Start the server using Docker (recommended):

docker compose up -d

This runs the MCP Datadog server and exposes health endpoints at:

• http://localhost:8081/health (Datadog logs/queries)

4. Verify the server is running:

curl http://localhost:8081/health

You should see Healthy.

Note: If you prefer running locally without Docker, you can run the .NET CLI option instead, but ensure Datadog keys are provided in environment variables as shown in the README.

Tips and common issues:

• Ensure the Datadog API Key and Application Key are valid and have the required scopes (Logs, Monitors, Metrics, Incidents, Dashboards).
• If you change the Datadog site (region), update Datadog__Site accordingly (e.g., datadoghq.eu or ap1.datadoghq.com).
• When using Claude Desktop/Code, configure the MCP server URL (http://localhost:8081) and verify the correct port based on your startup method (Docker uses 8081; .NET CLI uses 5100).
• If the server fails to start, check Docker logs with: docker compose logs -f, and ensure environment variables are loaded correctly from the .env file.
• The available tools include search_logs for Datadog Logs, list_monitors/get_monitor for Monitors, query_metrics/list_metrics/get_metric_metadata for Metrics, list_incidents/get_incident for Incidents, and list_dashboards/get_dashboard for Dashboards.
• For production deployments, consider exposing the server behind a reverse proxy and enabling TLS in front of http://localhost:8081 if needed.