Analyzes a single trace as a top‑to‑bottom waterfall of spans in time order. Split layout: analysis tabs on the left, the Span Detail panel on the right.


Get here: open a trace from Traces Explorer, Distributed Tracing, or View Traces.

Across the top you get the trace’s summary and quick actions: its Status (e.g. Completed), the Application it belongs to, the Trace ID, the Start Time, and the total Trace Duration — plus Share (a permalink back to this exact trace) and View Related Logs.

The Timeline tab

  • A flame‑chart overview at the top (see Flame Graph View), then the category legend, then the span rows.
  • Each row: a child‑count badge, operation / service, a duration bar colored by transaction category (e.g. Programming Language, HTTP, RPC Systems), the span’s Duration and % of Trace labels, indentation for hierarchy, and badges (Error, Missing Parent, Fast Calls). Controls: Expand All / Collapse All, full‑screen.

Reading a span bar

  • Light (tinted) fill = full span duration in its category color.
  • Opaque (solid) segments = the span’s exclusive (self) time — work not handed off to a child in a different service. Labels beneath: Duration, % of Trace, Exclusive Duration.
  • A large Duration but small Exclusive Duration ⇒ the span was mostly waiting on downstream; a large Exclusive Duration ⇒ the span itself is the bottleneck.

How to debug

  1. Follow the longest / errored bar down the tree to the bottleneck or failure point.
  2. Cross‑check the Latency tab (per‑service breakdown) for the dominant service.
  3. Select the span → inspect it in Span Details.