datadog · datadog-labs — GitHub Copilot plugin: features and installation
Connect GitHub Copilot to Datadog through an MCP server and ask about logs, metrics, traces, dashboards, or monitors in plain language.
copilot plugin install datadog@awesome-copilotWhat datadog does
The plugin gives GitHub Copilot access to Datadog data through a preconfigured MCP server. Ask about logs, metrics, traces, dashboards, and monitors. You need a Datadog account and either Copilot CLI or Visual Studio Code with the GitHub Copilot extension.
After installation, run /ddsetup to choose the Datadog domain and connect your account. Browser-based OAuth handles authentication by default; API and application keys are another option, set with DD_API_KEY and DD_APPLICATION_KEY. If Datadog MCP is already registered separately, such as in .vscode/mcp.json, disable that registration to avoid a conflict.
What it helps with
- Find error logs for a service over a chosen period
- Check which monitors are alerting
- Locate traces with high latency
- List Datadog dashboards
Example prompts for the agent
- Show error logs for
checkoutfrom the last hour. - Which Datadog monitors are alerting right now?
- Find
api-gatewaytraces with latency above 500 ms.
How to install datadog
Command from the source
copilot plugin install datadog@awesome-copilotStep by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own
Questions about datadog
How do I install datadog in GitHub Copilot?
In Copilot CLI, run copilot plugin install datadog@awesome-copilot. Then run /ddsetup to select your Datadog domain and connect your account.
What does the datadog plugin do?
It connects GitHub Copilot to Datadog through an MCP server, letting you query logs, metrics, traces, dashboards, and monitors.
What do I need to connect Datadog?
You need a Datadog account and either Copilot CLI or Visual Studio Code with the GitHub Copilot extension. Run /ddsetup to configure access; disable a separately registered Datadog MCP server to avoid a conflict.
Source text
SKILL.md · Apache-2.0
# Datadog Plugin for Copilot & VS Code Query your Datadog data directly from Copilot and VS Code using natural language. Ask about logs, metrics, traces, dashboards, monitors, and more. ## What you need - A [Datadog](https://www.datadoghq.com/) account - [Visual Studio Code](https://code.visualstudio.com/) (latest version recommended) with the [GitHub Copilot](https://marketplace.visualstudio.com/items?itemName=GitHub.copilot) extension, and/or the [Copilot CLI](https://github.com/features/copilot/cli) ## Getting started To install using the Copilot CLI: - Type `copilot plugin install datadog@awesome-copilot` To install using VS Code: 1. Enter `@agentPlugins datadog` in the search field of the Extensions activity. 2. Select the Datadog plugin and click **Install**. After installation: - Run `/ddsetup` in 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. ### Notes If you already have the Datadog MCP server registered separately (for example, in `.vscode/mcp.json`), disable or remove it to avoid conflicts. ## Using the plugin 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 ``` ## Can't connect? **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. ## Changing settings 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 ## Advanced usage ### Key authentication Instead of OAuth, you can authenticate using a Datadog API key and application key. 1. Run `/ddsetup` as usual to configure the MCP domain. 2. Set the following environment variables before starting Copilot / VS Code: ```bash export DD_API_KEY=your-api-key export DD_APPLICATION_KEY=your-application-key ``` 3. 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. ## Good to know - By default, authentication is handled via OAuth in your browser. Key authentication is also [supported](#key-authentication). - No Datadog credentials are sent to the AI model provider. ## Support - [Datadog MCP Server Documentation](https://docs.datadoghq.com/mcp_server/) ## Legal See the [LICENSE](LICENSE) and [NOTICE](NOTICE) files included with this plugin. For details on how Datadog handles your data, see the [Datadog Privacy Policy](https://www.datadoghq.com/legal/privacy).
© Datadog · Apache-2.0 · datadog-labs/copilot-plugin
Similar plugins
All plugins for DevOps and infrastructure →datadog by other authors comes first.