Skip to main content
Tracing

VoltOps Tracing

Tracing shows how a single user request runs from start to finish. As a developer, you use it to answer:

  • Why did the agent respond this way?
  • Which step failed?
  • Where did latency or cost spike?


When you select a trace from the list, you can open the Waterfall view or the Node-Based view and then dig into Logs and Feedback.

Trace Filters​

Filters help you find the right trace fast.

tip

For example use:

  • Status + Duration to see where a failure started,
  • Token usage + Cost to catch expensive runs, and
  • User ID + Conversation ID to inspect a specific user flow.

Trace filters overview


  • Status: isolate failed or retry-heavy runs.
  • Agent ID / Entity type: narrow down which agent or step is involved.
  • Token usage / Cost: find expensive or abnormal runs.
  • Duration: spot slow traces and latency outliers.
  • Feedback source / key: review feedback by source.
  • User ID / Conversation ID: drill into one user journey.

Table of Contents