Incorrect statistics for crashed state.

[beyondkmp]

Is there an existing issue for this?

• I have checked for existing issues https://github.com/getsentry/sentry-javascript/issues
• I have reviewed the documentation https://docs.sentry.io/
• I am using the latest SDK release https://github.com/getsentry/sentry-javascript/releases

How do you use Sentry?

Sentry Saas (sentry.io)

Electron SDK Version

5.11.0

Electron Version

34.0.0

What platform are you using?

Windows

Link to Sentry event

https://jupiter-ct.sentry.io/issues/6104662640/events/b6c61dca18e84a618af8e7be368b7657/?project=4508278211739648

No response

Steps to Reproduce

1. start
2. generate network error

Expected Result

Do not include it in the crashed statistics.

Actual Result

Include it in the crashed statistics.

This is caused getsentry/sentry-javascript#15072.

Currently, any event with level "fatal" is marked as crashed. However, this type of network error was not previously counted as crashed. After this change, it started being included.
I think we should not only check if the level is "fatal", but also verify whether the attachment's type is "event.minidump". Only when both conditions are true should it be considered as crashed.

Alternatively, you can fix it by checking: if exceptions exist and all of them are handled, then crash should be set to false.

[timfish]

Do you know where this event comes from and why it is set as fatal? Sentry classifies fatal as events that will end the process. I've checked the Sentry code and I can only see uses of fatal for unhandled errors and native crashes.