Dynamic atoms creation
kxepal opened this issue · comments
I wonder what is the reason for Folsom to having dynamic atoms creation for pid_port_fun_to_atom?
The origin commit says better string hinting and ip parsing , but I wonder is this is more important than having your Erlang app crush due to atoms table exhausting.
Yes, I'd read the README and saw the caution:
folsom_vm_metrics:get_process_info(). %% use with caution
folsom_vm_metrics:get_port_info(). %% use with caution
but this isn't serious since it make pointless providing metrics which you shouldn't use with all cost.
IIRC the caution with regards to those functions was due to locking that may occur in getting the underlying info from the ports and processes not the creation of atoms. I can't find details on that right now.
If you are running into atom exhaustion issues with running these functions I would entertain a PR that changed this back to lists with tests and etc.
Folsom has moved, please resubmit your issue at https://github.com/folsom-project Thanks!