Connect OpenClaw to Datadog
Let an OpenClaw agent search logs, query metrics, check monitors and manage downtimes in Datadog, with no credentials handed to the agent.
Datadog is the observability platform where engineering teams keep their metrics, logs, traces, monitors and dashboards in one place. It is also where a lot of on-call time goes: deciding whether an alert is real, searching logs across three services to find the first error, muting a monitor before a deploy and remembering to unmute it, and rebuilding the same dashboard panel for a new service.
With Claworc’s External Connections, an OpenClaw agent can work
inside your Datadog account on your behalf. You connect Datadog once through a popup,
Composio holds the resulting credentials, and the agent gets an auto-generated
claworc-datadog skill that tells it which Datadog tools exist and how to call them. The
agent never sees an API key, an application key or an OAuth token.
Tools available to OpenClaw agents
Section titled “Tools available to OpenClaw agents”Composio’s Datadog toolkit exposes 42 actions across logs, metrics, traces, monitors, dashboards, SLOs, incidents and synthetics. In practice an OpenClaw agent leans on a handful of capabilities:
- Search logs — query logs with Datadog’s search syntax over a time window and get back matching lines.
- Query metrics — pull a timeseries for any metric, such as p95 latency or error rate for one service.
- Check monitors — see which monitors are alerting and read the details of one.
- Mute and schedule downtime — silence alerts for a planned change, then bring them back.
- Search traces — find slow or failing requests in APM by service, resource or tag.
- List incidents — see what is open and who is on it.
- Post events — add a deploy or change marker so it shows up on graphs.
Every tool is called through a small broker running next to the agent, so the agent only needs to describe what it wants done in plain language and let the skill handle the rest.
Use cases
Section titled “Use cases”First-pass triage on a page
Section titled “First-pass triage on a page”When a monitor fires, ask the agent to “look at the checkout-api alert and tell me what changed.” It reads the monitor, pulls the relevant metric, searches the logs for errors in the same window and the traces for the slowest requests, then summarizes: when it started, what the errors say, and which endpoint is affected.
Deploy hygiene
Section titled “Deploy hygiene”Before a release, tell the agent to mute the monitors for the affected services or schedule a downtime window, and post a deploy marker. When the release is done, it unmutes and confirms every monitor is back to normal. No forgotten mutes, no silent hours.
Weekly reliability summary
Section titled “Weekly reliability summary”Every Monday, have the agent list the monitors that alerted last week, pull error-rate and latency trends for your key services, and check the incident list for anything still open. It posts a short report so the team meeting starts with numbers.
Get started
Section titled “Get started”Connecting Datadog takes three steps: an admin adds a Composio API key to Claworc’s
settings, you open the instance’s Settings tab, click + Add connection, pick
Datadog and sign in. The claworc-datadog skill appears in the instance a moment later.
The full walkthrough, including the exact Composio permissions the key needs, is in the
External Connections documentation.