New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
A user reported one bad reply, and I have no way to tell which of my agent's several tool calls produced it.
What you'll have at the end
A timeline of every model call, tool call, and retrieval step in that one request, with the exact step that went wrong marked
You need
A tracing setup that already records one shared id across every model call and tool call in a request, through an observability tool or your own logs, plus a report naming roughly when the bad reply happened.
Not covered
Wiring up tracing in the first place, and fixing whichever fault class the trace turns up, are both separate jobs from reading one trace to find the cause.
Leans on
Log every tool call so you can find what broke a run
If your agent has no per-request trace yet, that recipe builds the shared-id logging this one assumes already exists.
See your LLM's cost, latency, and errors on one dashboard
For a pattern across many requests instead of one reported reply, the dashboard view is the better starting point.
Checked 18 Aug 2026
Part of the LLMOps cookbook