I have a Python process that should not die that is multiprocessed. I have a queue process that feeds the queue every so often, but my program should use up queue items faster than the queue adds them. I have four worker threads that make requests to an API and update an SQL table with the results. However, if I use nohup and run it, it crashes. Even if I directly run it, it crashes. It used to run forever when I loaded it in PyCharm, but even that crashes it.

Thanks for the report.  This appears to be another instance of a long-standing unresolved problem when using networks functions in a thread application on macOS that invoke Python's _scproxy helper module to obtain possible network proxy configuration info from macOS. For more details see, for example, Issue13829. Assuming your application does not need to connect through a network proxy, suggest trying the workaround in https://bugs.python.org/issue30385#msg293958 which avoids the call to _scproxy by defining the no_proxy environment variable.

Thank you. Closing this issue.