This test has been observed failing intermittently in the CI.

It has several potential sources of instability:

- The test runs in AgentVM mode
- Receiver binds to the wildcard, sender sends to getLocalHost()
  (using wildcard/localhost instead of loopback/loopback has been a source of instability in the past)
- Sockets are not closed properly if an exception happens
- Publication of instance variables is not always Multi Thread safe
  (would require volatile or other means of synchronization)
- If the sender fails to send, the test will get wedged forever.
- Possibly better exception logging