Agent Tracing on the Claude Platform (Public Beta)

Send a step-by-step record of what your Claude agents did to the monitoring tools your team already uses, for developers on the Claude Platform.

Records every agent turn, model call, and tool use with how long it took, how many tokens it used, and its ID
Covers Claude Managed Agents, the Messages API, and the Claude SDKs; calls to connected tool servers are traced too
For Managed Agents and the Messages API you add a destination in the Claude Console. The SDKs send traces automatically
Uses an open industry format, so it works with any monitoring tool that accepts it. Named partners include Grafana Cloud, Honeycomb, Sentry, Splunk, Datadog, New Relic, PostHog, Langfuse, LangSmith, Braintrust, and Arize AI

Seeing why an agent was slow, expensive, or went wrong, one step at a time.

You have Claude agents running in production and need to debug or report on them in an existing dashboard.

Engineering and ops teams who already watch their systems in tools like Datadog, Grafana, or Sentry.

Staying informed is great. Staying ahead is better.

Pick a topic and leave with something you created.

Get new module drops and weekly AI strategy.