Please don't assume that maybe_stuck/1 is particularly smart: all it does is listing processes
that don't have their stack trace changed in N seconds or after N attempts. This is not necessarily
an indication of an issue.
There are processes in any Erlang system that do nothing but wait for and monitor other processes. They
can be supervisor processes or something else.
Specifically for gen:do_call/4, those calls can be different. Even if they are not, they are part of
[TLS] connection termination, which will be cleaned up by the heartbeat mechanism even if one of
the calls gets no response (e.g. because target process is dead or on an unavailable node).