Query your Datadog data directly from Copilot and VS Code using natural language. Ask about logs, metrics, traces, dashboards, monitors, and more.
- A Datadog account
- Visual Studio Code (latest version recommended) with the GitHub Copilot extension, and/or the Copilot CLI
To install using the Copilot CLI:
- Type
copilot plugin install datadog@awesome-copilot
To install using VS Code:
- Enter
@agentPlugins datadogin the search field of the Extensions activity. - Select the Datadog plugin and click Install.
After installation:
- Run
/ddsetupin Copilot to connect the plugin to your Datadog account. The agent will guide you through selecting the correct Datadog MCP domain. You can run this again at any time.
If you already have the Datadog MCP server registered separately (for example, in .vscode/mcp.json), disable or remove it to avoid conflicts.
Once connected, just ask the agent anything about your Datadog data:
Show me error logs for the "checkout" service from the last hour
What monitors are currently alerting?
Find traces for service "api-gateway" with latency > 500ms
List my dashboards
Never connected before? Run the /ddsetup command in Copilot. It will help you provide the correct Datadog MCP domain and set up the MCP server.
Was working before but stopped? Run the /ddconfig command in Copilot. It will check your site, authentication status, and network access to help diagnose the issue.
The plugin provides a few commands you can run in the agent to manage configuration:
/ddconfig— change your Datadog site or switch organizations/ddtoolsets— enable or disable groups of tools
Instead of OAuth, you can authenticate using a Datadog API key and application key.
-
Run
/ddsetupas usual to configure the MCP domain. -
Set the following environment variables before starting Copilot / VS Code:
export DD_API_KEY=your-api-key export DD_APPLICATION_KEY=your-application-key
-
Restart Copilot / VS Code (or reload the MCP server via
MCP: List Servers→ Restart Server). The server will use the API keys instead of prompting for OAuth.
On macOS, launch VS Code from the terminal after exporting the variables, or use a launcher that inherits your shell environment.
- By default, authentication is handled via OAuth in your browser. Key authentication is also supported.
- No Datadog credentials are sent to the AI model provider.
See the LICENSE and NOTICE files included with this plugin.
For details on how Datadog handles your data, see the Datadog Privacy Policy.