Datadog MCP Server

mcp-datadog-server is an MCP server that connects large language model clients to the Datadog API platform. Built with the official TypeScript SDK, it exposes a comprehensive suite of 351 tools, combining curated operations, standard CRUD functionality, and schema-generated endpoints. DevOps engineers, site reliability teams, and infrastructure operators use it to query observability data and automate monitoring resources directly through natural language interfaces. The server enables users to perform full lifecycle management for monitors, dashboards, downtimes, user accounts, and teams. It supports real-time telemetry retrieval, such as searching log archives and running metric timeseries queries, as well as managing synthetic tests and tracking operational incidents. With automated schema discovery and Zod validation, LLMs can autonomously inspect parameter requirements and format valid payloads. Built-in features such as request timeouts, retry handling, progress tracking, and site selection ensure dependable communication across Datadog regions.

Category: Cloud & Infrastructure

Tags: alerts, datadog, logs, metrics, monitoring

Visit Datadog MCP Server

How to install and configure Datadog MCP Server

1. Ensure Node.js 18 or higher is installed on your machine. 2. Obtain your Datadog API Key and Application Key from your Datadog account dashboard. 3. Install the server package globally via npm: npm install -g mcp-datadog-server 4. Add the server entry to your MCP client configuration file (such as Claude Desktop's claude_desktop_config.json): json { "mcpServers": { "datadog": { "command": "mcp-datadog-server", "args": ["serve"], "env": { "DD_API_KEY": "your_api_key", "DD_APP_KEY": "your_app_key", "DD_SITE": "datadoghq.com" } } } } 5. Save the configuration file and restart your MCP client.

What you can do with Datadog MCP Server

• Managing alerts by creating, reading, updating, or deleting Datadog metric, log, or composite monitors. - Querying timeseries metrics across infrastructure components and investigating active threshold breaches using natural language. - Executing advanced log searches with custom filters to inspect operational errors and audit traces. - Scheduling and canceling maintenance downtimes for specific hosts or scopes during standard deployment cycles. - Administering user accounts and team structures within Datadog directly through prompt-driven tool calls.

Key facts

• https://github.com/ClaudioLazaro/mcp-datadog-server
• Cloud & Infrastructure, Monitoring & Observability
• alerts, datadog, logs, metrics, monitoring

Part of MCP Servers

What can mcp-datadog-server do?

mcp-datadog-server provides 351 tools for interacting with Datadog APIs. It offers full CRUD capabilities for monitors, dashboards, downtimes, users, and teams. It also features curated tools for querying metrics, searching logs, managing operational incidents, and configuring synthetic tests through connected AI models.

How do I install mcp-datadog-server?

You can install mcp-datadog-server globally using npm by running npm install -g mcp-datadog-server. Alternatively, clone the repository locally and run npm install, or execute it directly without prior installation using npx via your MCP client configuration.

Which Datadog sites are supported?

The server supports multiple Datadog regions by configuring the DD_SITE environment variable. Compatible endpoints include datadoghq.com (US1), datadoghq.eu (EU), us3.datadoghq.com (US3), us5.datadoghq.com (US5), ap1.datadoghq.com (AP1), and ddog-gov.com (US Gov).

Which MCP clients work with mcp-datadog-server?

mcp-datadog-server is compatible with any client supporting the standard Model Context Protocol, including Claude Desktop and Cursor. Setup requires adding the command, arguments, and required API credentials to the client MCP configuration JSON.

What is the license for mcp-datadog-server?

mcp-datadog-server is released as open-source software under the Apache 2.0 license. The source code repository and license details can be accessed directly on GitHub.