Skip to content
Investigate

Errors & investigation

Triage repeated failures and inspect the evidence from each occurrence.

Events, groups, and incidents

An error event is one captured occurrence. Errors groups related events so you can review a recurring failure together. An incident is the response record for an issue that needs ownership and coordination; it can be linked to error evidence.

Triage an error

  1. Open Errors and narrow the view to the environment or application you are investigating.

  2. Check recurrence, first and recent occurrences, and whether the error began around a release.

  3. Open an occurrence for its stack trace, request, breadcrumbs, and captured session.

  4. Assign an owner or link an incident when the issue needs coordinated response.

Follow the failing request

A replay shows the captured browser activity before a failure. Request details explain which endpoint failed. Trace IDs connect captured browser and backend events when propagation and middleware are configured. Source analysis is most useful when the node reports the full deployed commit.

Not every event has every type of evidence. A backend-only event may have no replay. Missing headers, masking, sampling, or an unavailable source revision can limit what you can inspect.

Record the outcome

Leave the cause, mitigation, and remaining work where the team can find them. Resolve the group or incident when the failure has been addressed and service behavior is stable. Resolving a record does not itself deploy a code change.