Observability lets you watch the agent that powers Enterprise Search at run time — its conversations, cost, and quality.
Overview
Observability groups five tabs: Sessions, Traces, Spending, Memories, and Evaluation. These are the same monitoring screens an AI agent has, all pointed at the single Enterprise Search agent — so there's no agent picker to work through first. One difference from a standalone AI agent: usage analytics live under Insights, not here.
What's in This Section
| Topic | Covers |
| Sessions | Real conversations people have had with Enterprise Search. |
| Traces | The step-by-step execution behind a request — retrieval, tool calls, and model steps. |
| Spending | Cost tracking, monthly caps, and spending alerts for the Enterprise Search agent. |
| Memories | What the agent has remembered from past conversations. |
| Evaluation | Continuous scoring of the live agent's real sessions. |