Span and trace evals judge one model call or one request. Conversation-level evals judge the whole multi-turn session. Here is which agent failures each unit can and cannot see, and how to run both without double-counting.
Need help?
Contact usSpan and trace evals judge one model call or one request. Conversation-level evals judge the whole multi-turn session. Here is which agent failures each unit can and cannot see, and how to run both without double-counting.