datadog — Claude plugin: Datadog telemetry in chat
This plugin wires Datadog telemetry into Claude and Claude, so you can ask about production services in plain language, spot issues and fix them without leaving the chat.
Open on GitHubWhat datadog does
The plugin gives the agent access to your Datadog telemetry. You can ask about production applications, check what is happening with critical services, see the picture on a chart and work through the fix. All in natural language, no separate console.
There is a second mode too: agentic loops. The agent runs checks on its own so observability and service management posture stay in shape. That is a recurring pass rather than a one-off question.
One hard limit: the plugin is available to US1 customers only. If your tenant sits in another region, it will not work, and that is worth confirming before you build it into a process.
What it helps with
- check production service health from the chat
- trace an incident back through telemetry
- chart metrics without opening a dashboard
- remediate an issue in a critical service
- keep observability under regular agent checks
Example prompts for the agent
- show me errors in the checkout service over the last hour
- what is happening with payments latency?
- run an observability check on our critical services
Questions about datadog
How do I install the datadog plugin in Claude?
It comes from the plugin marketplace. The exact command order depends on how plugins are wired in your setup, so follow the plugin installation instructions for Claude.
What does the datadog plugin do?
It connects Datadog telemetry to Claude and Claude so you can ask about applications, investigate incidents and fix service issues right in the chat.
Who can use the datadog plugin?
US1 customers only. Tenants in other regions are out of scope.
