How are panic crashes usually logged?

Question

go

Prompt good practices on how to pledge raw panic.

Answer 1

If we talk about the http server - I adopted the practice from Django.
I have an http middleware that:
- calls recover
- responds normally to the request, but does not crash, i.e. 500 + what we already know about the error and are working on fixing it
- collects all information on request, http headers, cookies, server settings, environment variables, etc. and notifies me by mail / telegram
This gives me the opportunity to find out about the problem before someone writes to support.
People, in a good way, are surprised when something doesn't work out for them, the server answers 500, they don't understand what to do. And I already found out about it, corrected the mistake and wrote them a letter saying that there was a mistake, but everything has already been fixed :)

Answer 2

panic - abnormal termination of the program. It contains all the necessary information. We just leave the trace in its original form (you can't handle some panics) and the server is rebooted using reflex.
Clarifying questions in the comments. it is not entirely clear what is meant.

Answer 3

Sentry