Installation

Choose your preferred installation method and IDE below.

Get Your Datadog Keys

Before configuring any IDE, you need both a Datadog API key and an Application key:

1. Go to Organization Settings > API Keys
2. Create or copy an API key
3. Go to Organization Settings > Application Keys
4. Create or copy an Application key with read-only scopes

caution

Never commit your Datadog keys to version control. Always use environment variables or IDE-specific secret management.

Environment Variables

Variable	Required	Description
DD_API_KEY	Yes	Datadog API key
DD_APP_KEY	Yes	Datadog Application key
DD_SITE	No	Datadog site/region (default: datadoghq.com)

Supported Datadog Sites

Site	DD_SITE Value	Region
US1 (default)	datadoghq.com	United States
US3	us3.datadoghq.com	United States
US5	us5.datadoghq.com	United States
EU	datadoghq.eu	Europe
AP1	ap1.datadoghq.com	Asia Pacific
US1-FED (Gov)	ddog-gov.com	US Government

info

If your Datadog organization is not in the default US1 region, you must set DD_SITE to match where your keys were created. Using keys from one region against a different site will result in authentication errors.

---

Option A: Via npx (Recommended)

No installation needed. Your IDE will run the server on demand:

npx -y @vineethnkrishnan/datadog-mcp

This is the simplest method — just reference it in your IDE config below.

Option B: Global Install

npm install -g @vineethnkrishnan/datadog-mcp

Then use datadog-mcp as the command in your IDE config (instead of npx).

Option C: From Source

git clone https://github.com/vineethkrishnan/mcp-pool.git
cd mcp-pool
npm install
npm run build

Then use node /absolute/path/to/mcp-pool/packages/datadog/build/index.js as the command.

---

IDE Configuration

Claude Desktop

Config file location:

OS	Path
macOS	~/Library/Application Support/Claude/claude_desktop_config.json
Windows	%APPDATA%\Claude\claude_desktop_config.json
Linux	~/.config/Claude/claude_desktop_config.json

Quick access: Settings > Developer > Edit Config

Using npx:

{
  "mcpServers": {
    "datadog": {
      "command": "npx",
      "args": ["-y", "@vineethnkrishnan/datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

Using global install:

{
  "mcpServers": {
    "datadog": {
      "command": "datadog-mcp",
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

From source:

{
  "mcpServers": {
    "datadog": {
      "command": "node",
      "args": ["/absolute/path/to/mcp-pool/packages/datadog/build/index.js"],
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

tip

Restart Claude Desktop completely after editing the config. Changes are not picked up automatically.

---

VS Code (GitHub Copilot)

Config file location:

Scope	Path
Workspace	.vscode/mcp.json (in your project root)
User (macOS)	~/Library/Application Support/Code/User/mcp.json
User (Windows)	%APPDATA%\Code\User\mcp.json
User (Linux)	~/.config/Code/User/mcp.json

Quick access: Command Palette (Cmd+Shift+P) > MCP: Open User Configuration

warning

VS Code uses "servers" as the root key, not "mcpServers".

Workspace config (.vscode/mcp.json):

{
  "servers": {
    "datadog": {
      "command": "npx",
      "args": ["-y", "@vineethnkrishnan/datadog-mcp"],
      "env": {
        "DD_API_KEY": "${input:dd-api-key}",
        "DD_APP_KEY": "${input:dd-app-key}",
        "DD_SITE": "datadoghq.com"
      }
    }
  },
  "inputs": [
    {
      "id": "dd-api-key",
      "description": "Datadog API Key",
      "type": "password"
    },
    {
      "id": "dd-app-key",
      "description": "Datadog Application Key",
      "type": "password"
    }
  ]
}

The inputs feature prompts you for the keys securely — they never get stored in the file.

User config (without inputs):

{
  "servers": {
    "datadog": {
      "command": "npx",
      "args": ["-y", "@vineethnkrishnan/datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

---

Cursor

Config file location:

Scope	Path
Global	~/.cursor/mcp.json
Project	.cursor/mcp.json (in your project root)

Quick access: Settings > Cursor Settings > MCP

{
  "mcpServers": {
    "datadog": {
      "command": "npx",
      "args": ["-y", "@vineethnkrishnan/datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

Project-level config (.cursor/mcp.json) takes precedence over global config.

---

Windsurf (Codeium)

Config file location:

OS	Path
macOS / Linux	~/.codeium/windsurf/mcp_config.json
Windows	%USERPROFILE%\.codeium\windsurf\mcp_config.json

Quick access: Settings > Advanced Settings > Cascade section

{
  "mcpServers": {
    "datadog": {
      "command": "npx",
      "args": ["-y", "@vineethnkrishnan/datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

tip

Make sure MCP is enabled in Windsurf: Settings > Advanced > Cascade > Enable MCP.

---

JetBrains IDEs (IntelliJ, WebStorm, PyCharm, etc.)

JetBrains IDEs (2025.2+) support MCP servers through the settings UI.

Setup:

1. Open Settings/Preferences > Tools > AI Assistant > MCP Servers
2. Click + Add to add a new server
3. Configure:
   • Name: datadog
   • Command: npx
   • Arguments: -y @vineethnkrishnan/datadog-mcp
   • Environment Variables: DD_API_KEY=your-api-key DD_APP_KEY=your-app-key DD_SITE=datadoghq.com
4. Click OK and restart the AI Assistant

Alternative — manual config file:

Create or edit ~/.config/JetBrains/mcp.json:

{
  "mcpServers": {
    "datadog": {
      "command": "npx",
      "args": ["-y", "@vineethnkrishnan/datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

---

Claude Code (CLI)

Option 1 — CLI command (recommended):

claude mcp add datadog \
  --scope user \
  -e DD_API_KEY=your-api-key \
  -e DD_APP_KEY=your-app-key \
  -e DD_SITE=datadoghq.com \
  -- npx -y @vineethnkrishnan/datadog-mcp

Scope options:

• --scope user — available in all your projects
• --scope local — current project only (default)
• --scope project — shared with team via .mcp.json

Option 2 — manual config:

Edit ~/.claude.json (user scope) or .mcp.json (project scope):

{
  "mcpServers": {
    "datadog": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@vineethnkrishnan/datadog-mcp"],
      "env": {
        "DD_API_KEY": "your-api-key",
        "DD_APP_KEY": "your-app-key",
        "DD_SITE": "datadoghq.com"
      }
    }
  }
}

Verify it's working:

claude mcp list

---

Verify Installation

After configuring your IDE, test the connection by asking your AI assistant:

"List my Datadog monitors."

If the server is working, you'll get a response with your monitors. If not, check:

1. Both keys are valid — API key and Application key are both required
2. Site matches your region — if you're not on US1, set DD_SITE to your region
3. Node.js >= 20 — run node --version
4. Config file syntax — validate your JSON (no trailing commas)
5. Restart your IDE — most IDEs require a full restart after config changes

Security Recommendations

• Create a dedicated Application key for AI tooling with minimal read-only scopes
• Use a scoped API key if your Datadog plan supports it
• Never share config files containing API keys
• Add config files with secrets to .gitignore
• Rotate keys periodically via Datadog Organization Settings