[NBS]: Create test for YDB actor system interconnect DNS resolve cache invalidation

[jkuradobery]

Recently on one of our cluster there was a connectivity issue between one of filestore-vhost pods and HIVE/Schemashard tablets.
The problem was with incorrect DNS resolve in the interconnect internals. (All nodes were accessible from each other, DNS was working fine though). Before that, there were maintenance operations performed on the node in question and DNS was unavailable for quite some time. Our hypothesis is that the interconnect's DNS resolver caches failed DNS query's response.
Also, the issue affected a single process on a host, which points to the application's code.
We need to write a test in order to reproduce this bug for NBS.