Datadog MCP Server

Official Datadog MCP server for logs, metrics, monitors, dashboards, and incident management directly in your AI client.

Description

The Datadog MCP Server is Datadog's official Model Context Protocol server for AI-powered observability workflows. It connects MCP-capable AI clients such as Claude Code, Cursor, OpenAI Codex, or Gemini CLI to the Datadog observability platform. The direct product link points to the official Datadog documentation; the open-source repository datadog-labs/mcp-server contains source code, examples, and the license. According to the provider, the server acts as a bridge between observability data in Datadog and AI agents that need structured access to logs, metrics, traces, monitors, dashboards, and incidents.

What is the Datadog MCP Server

The Datadog MCP Server allows AI agents to access observability data directly, without switching between an IDE and the Datadog UI. According to the provider, it is HIPAA-eligible and supports a wide range of MCP clients. It runs exclusively as a hosted remote server and is accessible at mcp.datadoghq.com. EU customers use mcp.datadoghq.eu instead. According to Datadog, the server does not send data to third-party AI providers; the connected AI client determines which Datadog data is forwarded to its model provider. This distinction matters in data-sensitive environments: the MCP Server itself only processes tool calls and returns results, not prompts or model reasoning.

Prerequisites

Using the Datadog MCP Server requires an active Datadog account with appropriate permissions and an MCP-capable AI client. According to the provider, the server applies the authenticated user's existing access controls: role-based access control (RBAC), data access control, and log restriction queries apply exactly as they do for direct API or UI access. The server cannot grant a user access beyond their existing permission profile. Write operations require the corresponding permission, such as monitors_write, and are checked on each tool call.

Features: Logs, Metrics, Monitors, Dashboards, and Incidents

The core toolset includes, according to the provider, tools for logs, metrics, traces, dashboards, monitors, incidents, hosts, services, events, and notebooks. Additional toolsets can be enabled for specific areas: alerting for monitor validation and SLO search, dashboards for creating, updating, and deleting dashboards, DDSQL for structured queries over infrastructure resources and log data, error tracking, feature flags, Kubernetes resources, network monitoring, real user monitoring, synthetics, software delivery metrics, and workflow automation. The toolset system allows loading only the tools actually needed, preserving the AI client's context window. Using toolsets=all enables all generally available toolsets at once.

Setup

For Claude Code on the US1 site, a single command suffices: claude mcp add --transport http datadog https://mcp.datadoghq.com/v1/mcp. Many other clients can be configured through a .mcp.json file. The server uses Streamable HTTP as its primary transport and OAuth for authentication. Specific toolsets can be restricted using query parameters, for example ?toolsets=core,dashboards. Individual tools can be excluded with omit_tools without disabling the entire toolset. The Codex CLI uses different configuration parameters via HTTP headers.

Usage limits per the provider

According to the provider, the following fair-use limits apply: a burst limit of 50 requests per 10 seconds and 100,000 monthly tool calls. These limits can be adjusted on request. GovCloud is currently not compatible. Datadog states in its privacy notice that it collects certain usage data from the remote server, including interaction patterns, errors, and user identifiers, in accordance with the Datadog Privacy Policy. This data is retained for 120 days.

Security and data access

According to the provider, the server forwards the authenticated user's own credentials to Datadog APIs. Existing access controls apply automatically. For sensitive data categories such as logs or APM spans, Datadog recommends using data access control and log restriction queries to limit what the MCP Server returns. Because the AI client controls what is sent to its model provider, reducing data access also reduces potential exposure. Write tools require explicit permissions and are checked per call. A read-only user cannot execute write-enabled tools.

FAQ

Is the Datadog MCP Server free? The server itself is open source (MIT license) and access is available with an existing Datadog account. Current plan and usage terms are published at datadoghq.com/pricing. Costs arise from Datadog products, resource usage, the selected plan, and the chosen AI client — not from the MCP Server itself.

Can I run the server locally? The official deployment is exclusively as a hosted remote server operated by Datadog. The repository contains agent integration examples but does not provide a standalone local server mode as an official product variant.

How many tools are available? The number depends on the enabled toolsets. The core toolset covers the most common use cases; toolsets=all loads all generally available toolsets. The complete tool reference is available in the Datadog documentation.

Requirements

Active Datadog account, appropriate permissions (RBAC), and an MCP-capable AI client such as Claude Code, Cursor, or Codex.

Installation instructions

Quick start

For Claude Code on US1: claude mcp add --transport http datadog https://mcp.datadoghq.com/v1/mcp. EU customers use mcp.datadoghq.eu. Configure other clients via .mcp.json or client-specific instructions.

bash
claude mcp add --transport http datadog https://mcp.datadoghq.com/v1/mcp

What the Datadog MCP Server is

According to Datadog, the Datadog MCP Server is an official remote server for the Model Context Protocol. MCP is a standard that lets AI clients call structured tools instead of only sending free-form text. In this case, the server connects an AI client to Datadog functionality for observability. Observability means using logs, metrics, traces, monitors, dashboards, and incidents to understand the state of applications and infrastructure.

Datadog’s documentation describes the server as a hosted service rather than a standalone local product mode. That matters for planning: you are not simply installing an internal Datadog program on your own machine; you are connecting your MCP-capable client to a Datadog endpoint. The public repository contains source code, examples, and license information, while setup and operation are described in the official documentation.

Step 1: Prerequisites and connection

You need a Datadog account, appropriate permissions, and an AI client that can speak MCP over the supported transport. This can feel unfamiliar at first because the agent is not independently logged into Datadog; the connection is established through an authentication flow. According to the provider, OAuth is supported. OAuth is a method that grants confirmed authorization without placing passwords directly into configuration files.

Before the first connection attempt, identify the Datadog site, the client being used, and any company-specific data-access requirements. Some regions or environments may use different endpoints or have additional restrictions. Setup should therefore not be copied blindly from a random example, but taken from the current Datadog instructions for the exact client. Also check whether proxies, single sign-on, browser approval, or company policy affect the authentication flow.

Toolsets and practical value

Datadog describes toolsets as groups of tools for specific task areas. A core toolset covers common queries, while additional toolsets may enable areas such as dashboards, error tracking, Kubernetes, synthetics, real user monitoring, or workflow automation. For beginners, the key point is that you do not need to expose everything at once. The fewer tools are active, the easier it is to understand behavior, and the smaller the surface for mistakes.

In practice, the server can help diagnose incidents more quickly. A developer might ask which services are affected, whether logs match an error pattern, which monitors fired, or which dashboards show relevant signals. The benefit does not come from the AI receiving extra rights; it comes from structured access to existing observability data. The quality of the answer still depends on data quality, permissions, time range, question wording, and model behavior.

Permissions, privacy, and security

According to Datadog, the authenticated user’s existing access controls apply. These include role-based permissions and data-access restrictions. This is a central security principle: the MCP Server should not create rights the signed-in person does not already have. Even so, an AI client with broad read permissions can retrieve sensitive information from logs, traces, or dashboards. Roles should therefore be as narrow as possible, and sensitive data should be limited inside Datadog before the agent sees it.

Be especially careful with write-capable tools. If a tool can change monitors, dashboards, or workflows, decide who may approve those actions and how changes will be reviewed before production use. For analysis tasks, a restricted read-only scenario is often enough. The data-flow boundary also matters: Datadog notes that the connected AI client determines which returned information is forwarded to that client’s model provider. Less tool access therefore means less potentially forwarded data.

Limits and best practices

The Datadog MCP Server makes observability data easier to use from agents, but it does not replace incident processes, SRE experience, or security approval. A model can misread relationships, choose the wrong time range, or miss important operational context. Critical actions should always be reviewed by accountable people. A good starting point is a small use case: one service, a bounded time range, and a user with the minimum necessary rights.

Document which client is connected, which toolsets are enabled, which roles apply, and which kinds of data may appear in prompts or model responses. Review regularly whether the connection is still needed. If an incident workflow emerges, define when the agent only explains, when it suggests, and when humans decide. That keeps the Datadog MCP Server a helpful analysis tool rather than an uncontrolled bridge into sensitive operational data.

claude mcp add --transport http datadog https://mcp.datadoghq.com/v1/mcp

Authentication

OAuth flow through Datadog login; alternatively API key depending on the client. Existing RBAC permissions are applied directly.

Required access permissions

Permissions follow the authenticated Datadog user; no access beyond existing permissions is possible. Write tools are checked on each call.

Transmitted or stored data

The server receives tool calls and returns results. Datadog data is not sent directly to AI providers; the client controls that.

Security risks

Broad permissions, enabled write tools, or missing data access control can expose sensitive observability data or trigger changes to monitors and dashboards.

License and costs

License
MIT
Cost
free

The server is MIT licensed. Costs depend on the Datadog account, resource usage, selected plan, and AI client. See datadoghq.com/pricing.

Alternatives

Not recorded yet.

At a glance

Provider
Datadog
Status
Official server
Deployment
Remote
Current version
Not recorded yet.
GitHub stars
48
Last reviewed
08.09.2026

Repository and documentation

Categories

Supported clients

Related guides

Guides and background related to this entry.

No related guides yet.