After ignoring a group when a new error happens it restores the ENTIRE group... expected behaviour?
leroy30
Posted on
Oct 08 2014
It looks like after I ignore a group then as soon as a single error of that type happens again then it restores the entire group to the dashboard.
I think it would be better if it continued to ignore all errors up until that point; so I only had to deal with the new errors.
Is this by design, or is it a side-effect? Just wondering whether I should post a feature request?
jeremy
Raygun
Posted on
Oct 08 2014
Yes this is the intended behaviour. The status change applies to the group itself and not the occurrences contained within the group at the time.
If you ignored the existing occurrences would you still want to review the data for those at a later time? And would there be any condition that would trigger the group of occurrences that have been ignored to be considered active again?
Jeremy
leroy30
Posted on
Oct 08 2014
The issue is I get groups of errors with different causes. For example in my group for Exception: Exception of type 'Java.Lang.Exception' there are actually multiple stack traces within that group that I might address individually. I have one caused by NullPointerException within 'OnMeasure' and another caused by OutOfMemoryException somewhere else.
So if we fixed the out of memory exception or marked as ignored (because that specific device had issues we did not care to resolve) but then later the nullpointerexception occurrs, we might suddenly see 50 errors instead of just 1, when the other 49 should still be marked as ignored or resolved. We then have to check all those errors (probably forgetting the outofmemoryexception was caused by a dud device). It causes unnecessary rework.
Or another scenario -> that outmemoryexception was occurring but we better handled it. However now a broken device is causing this same issue even though we are handling it. This causes all 50 errors to come back, falsely indicating that the outofmemoryexception is happening for many users (assuming we forgot that we marked those as resolved or ignored).
It would be really great if 'ignore' or 'resolve' closed off those groups, and any new instances started a new group. I personally believe that is a more user-friendly way of handling it. And perhaps a 'show similar historical instances' button could be used for those that want to look back at the ignored exceptions.
I hope I'm making sense here!
leroy30
Posted on
Oct 24 2014
Here's a good example.
'Java.Lang.Exception' thrown in multiple locations, for different reasons. This happened today because of this exception:
android.view.WindowManager$BadTokenException: Unable to add window -- token android.os.BinderProxy@493b2fa0 is not valid; is your activity running?
And Raygun restored the entire group which includes resolved errors such as:
java.lang.IllegalStateException: Circular dependencies cannot exist in RelativeLayout
This is very misleading, and causes unnecessary stress when a whole group of errors is restored!