configurable cluster worker shutdown timeout

[azjezz]

currently, the context cluster worker has a hardcoded 5 seconds timeout on shutdown, i believe this aspect of the cluster should be configurable.

In the case of neu, the worker dispatches events when shutting down 1 2 3, allowing services to do clean up, which might take more than 5 secodns

[bwoebi]

Also, the ping-watcher and pong-replier in the Cluster class stops once the null is received. So, if the timeout is increased, as long as log messages are sent, it works fine. Once for 10 seconds no log message is logged, it's thus forcefully closed.