Pause/Resume improvements

[sobychacko]

          I think it would be more consistent and easier to follow if we instead do pause/resume exactly how start/stop is currently implemented. This would also remove the need to attempt to use the lifecycleMonitor in the subclass. 

Basically, we would have an abstract doPause()/doResume() where the concurrent does the 1-N delegation call to its child containers and each child container then does the work to pause/resume (as you have it now).

Originally posted by @onobc in #292 (comment)

[onobc]

Closed via 224eb0b