# How Sentry debugs its own AI agents

> Watch Ryan and Serge show how Sentry engineers use Agent Tracing to debug the agents behind Sentry's AI features and find failures across large volumes of production traces.

**URL:** https://sentry.io/resources/agent-tracing-series/

---

**Type:** Workshop | **Date:** ON-DEMAND

**Tags:** Workshop, Agent Tracing, AI

How does Sentry debug the agents behind its own AI features?

Ryan and Serge took a look at how Sentry engineers use Agent Tracing in production. They walked through how they collect traces, investigate agent behavior, and find the failures worth fixing when there are too many traces to review one by one, with real examples you can apply to your own agents.

This session covered:

* How Sentry instruments and debugs its own agents.
* How engineers use traces to investigate unexpected behavior.
* How the team uses signals, such as user frustration, to find problems at scale.
* How to get started with agent tracing and review growing volumes of traces.

In part one, [Debugging agents in different environments](/resources/debugging-agents-in-different-environments/), Serge covered the basics of traces and spans and followed agents running in an ecommerce store, Slack, and GitHub Actions. Watch it on demand if you're new to tracing.

## Related Resources

- debugging-agents-in-different-environments

- Seer, the Sentry MCP and CLI, or your own coding agent: where each one fits

- building-and-monitoring-ai-agents

---

*Source: [sentry.io/resources/agent-tracing-series/](https://sentry.io/resources/agent-tracing-series/)*