Issues

Every error your application reports is grouped into an issue: one row per distinct problem, however many times it has happened.

How grouping workslink

Events that share a fingerprint, normally the error type plus the top of the stack trace, collapse into a single issue. That is why a crash affecting thousands of users appears once, with a count, rather than flooding the list.

Minified or unsymbolicated stacks group poorly, because the frames differ from build to build. Uploading debug symbols and source maps improves grouping as well as readability. See Source Maps & Symbols.

Triagelink

  • Unresolved - needs attention.
  • Resolved - fixed. If it happens again on a later release, it reopens automatically.
  • Ignored - known and accepted. Stays out of the default view and stops alerting.

Select several issues to change their status together. The list also updates live, so a spike appears without reloading.

Filteringlink

Narrow by environment, release or status to answer questions like "what is new in production since this release". Filters are held in the URL, so a filtered view can be pasted to a colleague.

Reading one issuelink

Stack trace. Frames from your own code are highlighted ahead of framework and library frames, so the line that matters is the one you see first. If frames look like memory addresses or single-letter names, the matching symbols have not been uploaded yet.

Native crashes with no matching dSYM are held as "waiting for symbols" and re-processed automatically once you upload them. You do not need to reproduce the crash again.

Occurrences. Each occurrence keeps its own context: release, environment, device or runtime, the user it affected if your SDK sets one, and the breadcrumbs leading up to the error. Comparing two occurrences usually shows whether a problem is universal or specific to one platform or version.

Resolving. Mark an issue resolved once you have shipped a fix. If the same fingerprint arrives on a later release, BugWatch reopens it rather than silently starting a new issue. A regression is easier to spot as a reopened issue than as a new one.