Errors
The Errors tab groups exceptions into issues so a recurring failure appears as one item with a count and trend rather than thousands of separate events. The tab is available when your deployment has enabled the error rollup.
Issues List
Each issue shows:
| Column | Meaning |
|---|---|
| Issue | Exception type and representative message |
| Trend | Occurrence trend across the selected range |
| Count | Number of occurrences |
| Service | Service reporting the exception |
| Last seen | Most recent occurrence |
- Search by exception type, message, or operation.
- Click a sortable heading to change order.
- Use Load 50 more to retrieve another batch of grouped issues.
- Select an issue to open its resizable detail drawer.
Issue Detail
The drawer contains:
- Issue summary with service, span name, count, and last-seen time.
- An Occurrences chart across the full selected range.
- The latest available Stack trace, with a copy action.
- Recent occurrences with time, message, and trace ID.
The grouped count and chart come from the error rollup. Stack traces and recent occurrences are loaded from raw spans only after you select an issue.
Raw Occurrence Window
The Recent occurrences caption states the exact absolute interval queried. If the selected interval begins before the administrator's raw trace limit, APM moves the start to that retention floor. A completely expired historical range returns no raw events rather than reading older data.
Use Load 10 more to retrieve additional occurrences within that same allowed interval.
From Error to Trace
Select an occurrence to open traceX on its exact trace and failing span. This shows the request's other service calls, database work, events, and timing around the failure.
Create an Error Alert
When the issue is backed by the error rollup, the Occurrences chart menu can open a Grafana alert-rule draft scoped by the issue's numeric error key. The error message is not placed in the alert URL or labels. Review the query and threshold before saving; the action requires Grafana alerting permissions.
Use Cases
Triage a New Exception
- Sort the Issues list by last seen to surface what started recently.
- Open the issue and read the stack trace and its occurrence chart.
- Select a recent occurrence to see the failing span in its full trace.
Work Through the Loudest Errors First
- Sort by count over the range you care about.
- Check each issue's trend: a flat high count is background noise, a rising one is a regression.
- Use the service column to decide which team owns the fix.
Alert on an Issue You Cannot Fix Today
- Open the issue and use the Occurrences chart menu.
- Review the generated query and threshold in the Grafana alert draft.
- Save it so the issue's return is reported rather than rediscovered.
FAQ
How are exceptions grouped into issues?
An issue groups occurrences of the same exception type and representative message, so a failure that fired ten thousand times appears as one row with a count and a trend. The grouped count and the occurrences chart come from the error rollup.
Why is there no stack trace on an older issue?
Stack traces and recent occurrences are read from raw spans, which have a shorter retention limit than the rollup behind the count and chart. If the selected range starts before that limit, APM moves the query start to the retention floor, and a fully expired range returns no raw events at all.
Does an error alert include the exception message?
No. The alert draft is scoped by the issue's numeric error key, and the error message is not placed in the alert URL or its labels. Review the query and threshold before saving; the action needs Grafana alerting permissions.
Related Guides
- Getting Started - Scope, filters, and trace limits
- Services - The error rate an issue contributes to
- Traces - The request around a failing span
- logX - Logs from the same service