Management poller too slow for synchronous polling, should be disabled

The management poller has a timeout of 5000 ms deep in the app server client jar code. I haven't figured out how to isolate it yet, or to change the timeout value, and so until that can be done, it'd be better to ensure that the management poller is disabled for synchronous calls when the server is likely down (ie during a 'startup' call) until that can be figured out.

This would have the effect of the web port poller being used for such synchronous calls. This could be reversed once sufficient api was discovered / added to control the timeouts.