leviathan / getexceptional-api-wrapper Goto Github PK

Some unrelated exceptions are getting grouped together in Exceptional.

I spoke to Ciaran Lee at Exceptional about it and he noted that the problem lies with the method they use to group errors together, which depends on some data that this wrapper is not submitting.

These are the details that they are getting submitted via GTExceptional:

• The first line of the backtrace
• The first 15 chars of the error message
• The exception class

And these are the missing details:

• the controller name (a throwback to the rails specific origin of exceptional, this field is empty in your exceptions)
• the action name (a throwback to the rails specific origin of exceptional, this field is empty in your exceptions)

Without controller and action names, unrelated exceptions that nonetheless match based on the first three conditions are grouped together.

Do you think there is any way to use the controller and action fields to help avoid inappropriate grouping?

Some iOS exceptions seem to display odd dates. In some cases the "x minutes ago" even goes backward. Every time I refresh, X gets smaller. Maybe it has something to do with timezones? (mine is properly set in my account settings)

I posted this messages to the support site and Claran Lee replied with the following:

I'm guessing what has happened here is that you sent us an exception whose occurred_at timestamp was in the future (hence the x minutes ago was counting down initially). I'm not familiar with the iOS client code, but it might be worth checking whether it converts the time to UTC before it sends the error data to us.