# Debugging agents in different environments

> An ecommerce chatbot, a Slack agent, and a PR review agent can each fail across multiple steps. Watch Serge show how to use Sentry Agent Tracing to follow an agent run, inspect model and tool calls, and find where things went wrong.

**URL:** https://sentry.io/resources/debugging-agents-in-different-environments/

---

**Type:** Workshop | **Date:** On-Demand

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

**Presented by:** Sergiy Dybskiy

An ecommerce chatbot, a Slack agent, and a PR review agent run in different environments. Each can fail across multiple steps, making the final response only part of the story.

Watch Serge's introduction to tracing and how it applies to AI agents. Through live walkthroughs, you'll see how to follow an agent's execution, inspect model and tool calls, and find where things went wrong. No prior tracing experience is needed.

You'll learn how to:

* Read traces and spans, and understand the extra context agent tracing provides.
* Add agent tracing to an ecommerce chatbot, a custom Slack agent, and a PR review agent running in GitHub Actions.
* Inspect inputs, outputs, tool calls, and token usage to investigate errors and slow responses.

This is part one of a two-part series. In part two, [How Sentry debugs its own AI agents](/resources/agent-tracing-series/), Ryan and Serge show how Sentry engineers use Agent Tracing in production.

---

*Source: [sentry.io/resources/debugging-agents-in-different-environments/](https://sentry.io/resources/debugging-agents-in-different-environments/)*