Trace explorer

The Traces page (at /traces) searches the span store directly. Filter spans by service, operation, duration, status, or a query, read latency and error rate for what matches, and open the trace behind any span. It stays empty until spans are arriving. See Send traces, and check TRACE_INDEX_ID in Environment variables if spans should be there.

Filters

The toolbar filters every chart and table on the page.

Filter	Matches
Service	One service, by `service_name`.
Search spans	A query in Quickwit query syntax over span fields. Paste a trace ID to open that trace.
Duration	A preset band from < 10 ms to > 10 s, or Custom… with Min ms and Below ms.
Error / OK	Spans whose status is `error`, or every other span.
Root only	Each trace’s root span, one per trace.
Time range	Up to 30 days.

Root only with Error lists traces whose root span failed. Errors deeper in a trace don’t match. Queries run against the fields of the span store, for example:

# HTTP server errors
span_attributes.http.response.status_code:[500 TO 599]

# PostgreSQL calls from one namespace
span_attributes.db.system:postgresql AND resource_attributes.k8s.namespace.name:production

If a query doesn’t parse, the error shows under the toolbar and the results on screen stay those of the last good query.

Charts

Three charts cover every span the filters match:

Latency over time: p50, p95, and p99 span duration.
Error rate: error spans as a share of matching spans.
Requests: matching spans per interval.

Every matching span counts, whatever its kind, so one request crossing four services counts four times. Turn on Root only to count one span per trace. Service health counts only entry spans as requests. Drag across any chart to narrow the time range to that window.

Operations

The Operations tab groups matching spans by span name. Each row shows:

the services that emit it
span count, with a sparkline and the rate per minute
error rate and error count
p95 duration

Click an operation to filter the page to it and switch to Spans. The filter shows as a chip under the toolbar. If only one service emits the operation, the service filter is set too. Operation names here are raw span names. Service health names operations by route instead. See How Rootprint names an operation.

Spans

The Spans tab lists matching spans, newest first, 50 at a time. Each row shows the start time, service, operation, duration, status (with the HTTP status code when the span has one), and trace ID. Sort by Time or Duration. Click a row to open its trace with that span selected. See Read a trace.

Links and URL state

Every filter, the time range, the active tab, and the span sort are part of the URL, so you can bookmark or share a filtered view.

Limits

Limit	Behavior
200 operations	The span names with the most spans. Past that, the tab says it is showing the top 200.
100 services	The service picker lists up to 100 services seen in the range.
10,000 spans	Load more on the Spans tab stops at about 10,000 spans.
2,000-character query	A longer Search spans query is rejected.

Get Started

Send Logs

Traces

Search Logs

Manage Rootprint

Filters

Charts

Operations

Spans

Links and URL state

Limits

Next steps

Read a trace

Service health

​Filters

​Charts

​Operations

​Spans

​Links and URL state

​Limits

​Next steps

Read a trace

Service health

Filters

Charts

Operations

Spans

Links and URL state

Limits

Next steps