Skip to content

Traces

The Traces page is where you find and read distributed traces — the end-to-end record of a request as it moves across your services.

Searching

Filter traces by service, application, operation, duration, and status to narrow down to the requests you care about (the slow ones, the failures, a specific endpoint). Results page in as you scroll down (infinite scroll), and a scatter view plots traces so outliers stand out.

The trace panel

Clicking a row — or a dot in the scatter — opens the trace in a slide-in side panel rather than navigating away, so you keep your search list behind it. The panel is wide and resizable, and the open trace lives in the URL (?trace=<id>), so it's shareable and works with browser back/forward. A "Full page ↗" link opens the same trace as a standalone page.

Inside, you get summary cards and the waterfall — the spans of the trace laid out in time, showing where the request spent its time and where it failed.

Span detail

Click a span in the waterfall to open its own detail panel, stacked over the waterfall, with the span's timing and attributes.

Jump to logs

Both the trace header and a span's detail carry a "View logs for this trace/span" link. It opens the Logs page pre-filtered to that trace_id (and span_id for a span), so you can read exactly the log lines a request produced.

  • Logs — jump from a request to its logs and back.
  • Services — the entities a trace passes through.
  • Query Explorer — aggregate trace data with OQL.