Plugin restart after failure does not re-initialise Netty worker group

[praseodym]

This plugin has the same problem as described in logstash-plugins/logstash-input-beats#289, where the Netty worker group can end up in a dead state (e.g. because of a BindException caused by an in-use port) and the plugin is restarted without re-initialising the worker group. This will cause Logstash to be stuck in a meaningless plugin restart loop.