← Explore all mcp servers

OBSERVABILITY / Datadog

Datadog MCP

Query Datadog telemetry and operate permitted monitoring resources from an authenticated AI client.

“The Datadog MCP Server acts as a bridge between your observability data in Datadog and any AI agents”

01 / THE REASONING

Why this made the selection.

  • Connects logs, traces, metrics, incidents and monitoring context already held in Datadog.
  • Existing user permissions and configurable toolsets provide practical boundaries for an agent integration.
  • The overview describes a remote MCP server that lets clients such as Cursor, OpenAI Codex, and Claude Code query Datadog products including APM, logs, metrics, monitors, dashboards, and security signals.
  • Setup treats OAuth 2.0 as the usual login, separates tools that read data from tools that create or modify resources, and still requires the underlying Datadog resource permission for either.

02 / THE REVIEW RECORD

What we actually inspected.

Source review has boundaries.
A clear record is more useful than a “safe” badge.

Material inspected

  • Datadog MCP overview, setup instructions, tool reference and corresponding publisher documentation source.
  • https://docs.datadoghq.com/mcp_server/
  • https://docs.datadoghq.com/mcp_server/setup/
  • https://docs.datadoghq.com/mcp_server/tools/
  • Round 2 reread of acquired publisher setup and permission evidence; see immutable wave review.

Our findings

  • The main hosted server is distinct from Datadog’s local Code Security MCP server.
  • The setup guide supports toolset selection and omit_tools; these do not replace account-level authorization.
  • The overview presents one remote server for most Datadog products and tells readers to use a different, local-only Code Security MCP server for SAST, SCA, IaC, secrets, and SBOM scans.
  • That server answers individual tool calls, does not receive the prompt or the model's reasoning, returns results only to the calling client, and makes no outbound calls to external domains. Most listed tools query Datadog backends with no model involved; a few use Datadog-hosted models for semantic search or natural-language queries.
  • Permissions are the user's own: reading needs the read MCP permission plus the resource permission, writing needs the write MCP permission plus the resource permission, and the server cannot expose anything that user cannot already see in the Datadog UI. Admins can hide write tools organization-wide and can refuse connections from networks outside an IP allowlist.
  • The tools reference groups each tool by toolset and permission and says the set is still changing. It also states things particular tools do not do, including incident timeline data on get_datadog_incident and mutating actions in the assistant toolset.
  • Toolsets can be selected with the endpoint query or X-Datadog-MCP-Toolsets header; the ChatGPT app preview has a narrower site availability than generic MCP.

Not established by this review

  • Server launch, authenticated tool calls, and release-to-source parity were not tested.
  • The resource was not executed and its tools were not called.

The review applies to the material and revision named here. A newer upstream release can change its behavior.

03 / PUT IT TO WORK

Connect a server deliberately.

Upstream setup instructions ↗
  1. Open the official setup guide and select your Datadog site to obtain its correct endpoint.
  2. Follow the client-specific HTTP/OAuth instructions and complete authorization.
  3. Have an administrator grant the needed MCP and resource permissions, then choose only relevant toolsets.

Before you start

  • A supported Datadog site and account.
  • MCP Read permission and relevant resource permissions; MCP Write plus resource write permission for mutations.

04 / FOLLOW THE EVIDENCE

The source trail.

Our notes are separate from the original resource.
Check upstream before adopting a new version.

  1. https://docs.datadoghq.com/mcp_server/

    Supports: summary, upstreamDescription, whySelected, bestFor, limitations, install, access, compatibility, transports, review

  2. https://docs.datadoghq.com/mcp_server/setup/

    Supports: install, access, compatibility, transports, limitations

  3. Datadog tool reference ↗Checked

    https://docs.datadoghq.com/mcp_server/tools/

    Supports: summary, bestFor, review, access

Evidence & Ask