MCP For Claude Marketplace
Datadog

Datadog MCP Connector for Claude

A+

Monitor infrastructure, APM and logs via Datadog — query metrics, audit monitors, search logs and track incidents from any AI agent.

16 tools Official Updated Jun 28, 2026 Official Vinkius Partner
More Details Connect to Claude

Connect your Datadog account to any AI agent and gain full observability over your entire infrastructure, applications and logs through natural conversation.

What you can do

  • Monitor Management — List, create, update, mute and unmute alert monitors across metric, anomaly, log, service check and synthetics types
  • Metrics Querying — Query raw metric timeseries data with Datadog's query syntax to analyze CPU, memory, custom business metrics and more
  • Log Search — Search structured and unstructured log events using the same query syntax as the Log Explorer, filtering by service, host, status and any indexed attribute
  • Dashboard Discovery — List all dashboards, view their widget configurations and audit shared access without opening the Datadog app
  • Synthetics & SLOs — Audit your synthetic test coverage and Service Level Objectives to track SLA compliance across teams
  • Incident Tracking — View active and recently resolved incidents with severity, responder assignments and postmortem status
  • Infrastructure Inventory — List all monitored hosts with their tags, metrics summary and agent version
  • Team & User Auditing — Review team membership, user roles and access permissions to maintain organizational security

How it works

  1. Subscribe to this server
  2. Enter your Datadog API Key and Application Key
  3. Start monitoring your stack from Claude, Cursor, or any MCP-compatible client

Stop context-switching to the Datadog dashboard every time an alert fires. Your AI acts as a dedicated SRE.

Who is this for?

  • SREs & On-Call Engineers — instantly triage monitors, search error logs and check incident status without opening Datadog
  • Engineering Managers — audit monitor coverage, review SLO compliance and track team ownership across services
  • Developers — query metrics, inspect log events and verify synthetics directly from your IDE
apminfrastructure-monitoringincident-responsemetrics-queryingalertingcloud-ops

16 tools expose this connector's capabilities to your AI agent.

create_monitor

Requires the monitor type (metric, anomaly, service check, event, log, process, rum, synthetics), a query string (e.g. "avg(last_5m):avg:system.cpu.user{host:myhost} > 80"), a notification message (using @user, @slack, @pagerduty) and a name. Optionally set tags, priority, renotify interval and threshold windows. Create a new Datadog monitor

list_dashboards

Use to discover available dashboards before opening a specific one. List all Datadog dashboards

get_dashboard

Provide the dashboard ID. Get details for a specific Datadog dashboard

get_monitor

Provide the numeric monitor ID. Get details for a specific Datadog monitor

list_hosts

Each host reports CPU, memory, disk, network metrics plus custom tags. Optionally filter by a tag string (e.g. "env:production") to narrow results. List hosts monitored by Datadog

list_incidents

Each incident has a title, severity, status (active, resolved), timeline, responder assignments and postmortem status. Use to audit ongoing incidents and review resolution patterns. List Datadog incident management records

list_monitors

Monitors track metrics, anomalies, service checks and events. Each monitor has a type (metric, anomaly, service check, event, log), name, query string, notification message and current status. Use this to audit your alerting coverage. List all Datadog monitors

mute_monitor

Useful during maintenance windows or known incidents. Provide the monitor ID. Optionally set an end timestamp for auto-unmute or a scope to mute only specific sub-alerts. Mute a Datadog monitor

query_metrics

The query string uses Datadog syntax like "avg:system.cpu.user{host:myhost}". Provide Unix timestamps for the from/to range. Useful for analyzing metric trends without opening a dashboard. Query Datadog metrics timeseries

search_logs

Supports filtering by source, service, status, host and any indexed attribute. Example query: "service:api status:error". Returns matching log entries with full context, host info and trace ID if available. Search Datadog logs

list_slos

Each SLO defines a target availability percentage (e.g. 99.9%) for a service over a time window (7d, 30d, 90d). Useful for auditing SLA compliance across teams. List Datadog Service Level Objectives

list_synthetics_tests

Each test has a type, target URL, status, locations and check frequency. Use to audit your synthetic test coverage and verify endpoints are being monitored. List Datadog Synthetics tests

list_teams

Teams group users for ownership of monitors, dashboards, SLOs and incidents. Each team has a name, handle, description and user membership list. List Datadog teams

unmute_monitor

Provide the monitor ID. Optionally set a scope to unmute only specific sub-alerts. Unmute a Datadog monitor

update_monitor

Provide the monitor ID and any fields to update: name, query, message, tags, priority or thresholds. Only the fields you provide will be changed. Update an existing Datadog monitor

list_users

Use to audit access, identify inactive accounts and verify user permissions. List Datadog users

See how to talk to your AI agent using Datadog.

Show me all monitors that are currently in alert state.

I've scanned your Datadog monitors and found 3 currently in Alert state: 'High CPU Usage on web01' (metric alert, triggered 12m ago), 'API Latency P99 > 500ms' (query alert, triggered 5m ago) and 'Database Connection Pool Exhausted' (service check, critical). Would you like me to check the recent logs for any of these?

Search for error logs from the payment-service in the last hour.

I found 14 error logs for payment-service. The most common pattern is a 'ConnectionTimeout: upstream service did not respond within 30s' from host prod-web-03, occurring 9 times in the last hour. Would you like the full log entry with stacktrace?

What's our API error rate over the past 24 hours?

I queried the metric `sum:api.http.5xx{env:production}.as_count()` for the past 24h. The error rate peaked at 142 errors/min at 14:30 UTC and has been trending down to ~15 errors/min currently. The 24h average is 38 errors/min. Would you like me to correlate this with recent deployments?

The **API Key** authenticates your requests to the Datadog platform and is required for all endpoints. The **Application Key** is an additional layer of authorization that controls what actions your integration can perform. Both are generated in Organization Settings > API and Application Keys. Most Datadog API endpoints require both keys.

Related Connectors

CC

Checkout Champ MCP

8 tools Official

Manage CRM and order workflows via Checkout Champ — track leads, monitor orders, and analyze transactions directly from any AI agent.

A+ View details →
GD

Google Drive MCP

15 tools Official

Manage your files — search, create, share, and organize your Google Drive via AI.

A+ View details →
EA

EZO Asset Intelligence MCP

10 tools Official

Equip your AI agent to manage fixed assets, track inventory, and monitor checkouts via the EZO.io (EZOfficeInventory) API.

A+ View details →
TH

ThoughtSpot MCP

7 tools

Search and analyze business data by interacting directly with your ThoughtSpot metadata and Liveboards via your AI agent.

A+ View details →