Race condition during node startup

Summary: Potentially, Play can start receiving and processing requests for this node before it has joined the Cluster; badness will presumably result.

ConductR turns out to still have an explicit signal for when to stop processing, it's just buried in the Play/ConductR library. See the ConductRLifecycleModule for an illustration of how it works by default, this Case for Chris' response to what we should be doing to handle it, and this new section of the documentation for the official approach.

Overall, that doesn't sound too hard once I have the infrastructure in place to be monitoring Cluster state.