The node monitor race in `net_kernel:get_status()` is present both in all (?) R14B releases as well as R15B. It's not going to be fixed anytime quickly. So, the work-around involves even more knowledge of the internals of `net_kernel`, but we were already breaking that barrier: in for a penny, in for a dollar. Passes basho_expect's regression_az1018 test.
Change the way busy port messages are filtered. Old way: log all events up to a count. New way: log all events for a unique port up to the count. Note that the new way still counts all events, regardless of uniqueness, to log the number of messages suppressed by the filter. In the case of busy_dist_port events, we attempt to find the Erlang node name that is on the far side of the port's TCP socket. This can help diagnose point-to-point network congestion problems.