Find the request that matters
Filter by service, minimum duration, and up to five span attributes and five resource attributes. A duration histogram shows how the slow ones are spread.
Platform
Follow one request through every service it touched.
Traces lists the requests that started in your time range and opens any of them as a waterfall of spans. You see which service was slow, which call failed, and the logs written during that span.
What you get
Read the waterfall
Search inside a trace, highlight its critical path, and open any span to see its attributes, events and links.
Land on the logs
From a span, open the first 50 log lines written during it, or all the lines for the trace. They are matched by trace ID and span ID, not by guesswork.
How it works
Your service passes context along
An OpenTelemetry SDK adds a trace ID to each request and carries it from service to service.
Spans arrive at your workspace
Each service sends its own spans. Onyx joins the ones that share a trace ID.
You open the trace
The waterfall shows every span in order, with the failed ones marked and the logs one click away.
Limits worth knowing
- Large traces are cut. A trace with more than 1,000 spans shows the first 1,000.
- Age matters. A trace older than your plan keeps data opens as Trace not found.
- Lists load in pages. You see 50 traces at a time and load more.
Open your first trace.
Send traces over OpenTelemetry and they appear in the list.