Sign in

Monitoring

Performance and traces

Spans from Sentry SDKs (tracesSampleRate) or OpenTelemetry appear on the Performance page. Server and consumer spans, and roots, are transactions. Sending is covered in Traces and performance under Send more; this page is about reading the result.

1Performance page

Per transaction over 24 hours or 7 days: count, throughput, p50, p95, p99 and error rate, top 100 transactions. An error is a span whose status is not ok or whose HTTP status is 500 or above.

2Transaction page

A duration histogram up to p99 and the 20 slowest traces.

3Trace page

/traces/<traceId> shows a waterfall of spans with details, the errors captured in the same trace, and an N+1 flag when one database statement repeats 10 or more times under one parent span. An issue whose event carries a trace id shows View trace.

Good to know

  • Spans per month: Free 1M, Solo 5M, Team 20M, Scale 100M; kept 7 days on Free and Solo, 30 on Team and Scale. Past the allowance spans are dropped, never billed.
  • Percentiles are computed on read. At millions of spans in a window the 7-day view slows down; roll-ups are planned.
  • Retried requests can store duplicate spans, which shows as slightly high counts.

Stuck? Email support@blipit.io. Keys and the exact DSN for each project are under API keys in app.blipit.io.