agentName). The
Agents view rolls every trace up by agent, independent of which workflow it
ran in, so you can answer “how is my retriever behaving overall?”
Set agentName on a call’s integration to populate this view.
Agents list
A card per agent over the selected range: span count (and the LLM-span subset), tokens, p95 latency, and cost, with an error badge when any of its spans failed. A toggle switches between this card view and a compact table.Agent detail
Opening an agent shows:- Stat cards — spans, error rate, p95 latency, and total cost.
- Spans & errors and Latency — per-bucket trend charts (spans with errored spans overlaid; p50 / p95 / p99 latency) over the selected range.
- Cost breakdown — a donut of cost per pricing dimension across the agent’s priced spans.
- Trace flow — a graph of the agent’s most recent trace’s steps and tool calls (model logos for LLM steps, tool icons for tools). Click through to the full trace.
- Recent traces — the agent’s latest runs with spans, tokens, duration, and cost; click a row to open the trace.