Sign in Book a demo
Back to Tools
Developer Tools API Key
Build Your Agent View Datadog Docs
datadog logo

Datadog MCP Tools for AI Agents

Let an agent search logs and APM spans, inspect a trace, discover usable facets, and aggregate events over a time window.

Available Tools

5 tools available in this integration

Datadog.AggregateEvents

Aggregate log events or spans over a window, optionally grouped and bucketed by time.

Datadog.DiscoverFacets

Probe which standard facets are groupable and filterable for a dataset in a window.

Datadog.GetTrace

Retrieve a single trace assembled with a summary, bounding the span list.

Datadog.SearchLogs

Search Datadog log events matching a query over a time window.

Datadog.SearchSpans

Search Datadog APM spans matching a query over a time window.

Working directly against the raw API? 588 more tools in the Datadog API server →

Why agent-optimized tools matter

Built for intent, not API schemas

Ask the agent to "find the traces behind checkout errors from the last hour and explain the slowest span." Arcade helps it use the right Datadog tools without guessing query syntax, facet names or trace IDs.

Up to 100x fewer tokens

Datadog searches return focused log and span records, while trace retrieval returns a summary with a bounded span list. The agent gets useful evidence without loading an entire dataset or unbounded trace into the model.

Complete the workflow end to end

Search logs, search spans, inspect available facets, aggregate events over time, and retrieve a bounded trace through one toolkit designed around the complete observability investigation task.

Make actions predictable and controlled

Arcade tools return structured results and useful errors across users and runs. Teams can apply scope restrictions, redaction, and approval policies so agents follow the same controls every time.

Securing Datadog

The more capabilities agents get, the more useful they are. However, more capabilities bring more risks of undesired effects such as data loss and exfiltration.

This is why Arcade provides you with contextual rules that dictate how and when a tool is allowed to run. Arcade’s contextual access lets you govern tool visibility and behavior at every stage of execution and control who can see a tool, what inputs are allowed, and what comes back.

Learn about Contextual Access Policies

For Datadog, you can implement some of these policies:

Access when tools are listed for a user
  • Decide which tools an agent is allowed to see at all. A tool that is never listed cannot be called, whoever is asking.
Pre-execution before each tool call
  • Inspect the inputs a call is about to run with, and block it or send it for human approval when they fall outside what the agent should do unsupervised.
Post-execution after the tool returns
  • Redact log message bodies, email addresses, access tokens, and authorization headers from Datadog.SearchLogs results before they reach the model.
  • Deny Datadog.GetTrace results when spans contain customer identifiers, session tokens, or request and response bodies covered by your data-handling policy.
  • Strip database statements, HTTP headers, and user identifiers from Datadog.SearchSpans results so private request data does not reach the model.

How to use Datadog with your agent

Pick your MCP client below and follow the setup to connect Datadog to your agent.

Your questions, answered

A Datadog MCP server exposes Datadog capabilities as structured tools that an AI agent can call through the Model Context Protocol. It can let an agent search logs and APM spans, aggregate events, inspect available facets, and retrieve individual traces.

Arcade provides a Datadog integration with agent-optimized actions, shared-key authentication, approval and redaction policies, and one MCP gateway for supported clients. It also brokers credentials and provides runtime controls around Datadog API calls.

Arcade agents can search Datadog logs and APM spans over a time window, discover which facets can be filtered or grouped, aggregate events or spans by time and facet, and retrieve a single trace with its summary and bounded span list. These actions are read-only; they do not create, update, or delete Datadog data.

The integration uses a shared Datadog API key or secret configured once. The key or secret must have the Datadog permissions needed for the datasets and read operations your agent will query; grant the narrowest access that supports the workflow.

No. Users do not each connect a separate Datadog account through this integration. A shared API key or secret is configured once, brokered by Arcade, and never handed to the model, so calls run with the access granted to that credential.

The Datadog actions Arcade ships are read-only, so they do not make consequential changes such as modifying monitors or deleting data. Contextual Access policies can still require human approval before a call, redact fields from its response, or hide the actions from an agent entirely; the call proceeds only after the policy is satisfied.

Arcade proxies Datadog API calls and does not persist your content as part of normal tool execution. The shared credential is brokered and stored encrypted at rest, and the model never receives it.

Any framework or client that can connect to Arcade's MCP gateway or tool APIs works, including LangChain, OpenAI Agents, CrewAI, Claude Desktop, Claude Code, Cursor, Visual Studio Code, Microsoft Copilot Studio, and GitHub Copilot.

Arcade refreshes access while the configured Datadog credential remains valid. If the key or secret is revoked, rotated, expires, or can no longer be used, calls stop until the credential is reconnected or updated.

Datadog's API rate limits and any limits on the requested time window or result set still apply when calls run through Arcade. Trace retrieval also bounds the span list returned for a single trace, so the agent receives a focused result rather than an unbounded trace payload.

Ready to actually ship your agent?

Connect your agent to Datadog with a shared API key, agent-optimized tools and approval controls for consequential actions.

Build Your Agent Read the Datadog Quickstart