(stable) When daemons hibernate, mark them as not busy

The idle timer was not being reset on a hibernation exit vs a scaledown exit,
so the pool could get sort of "stuck" relaunching the second daemon. Instead,
reset the idle timer on hibernation exit.

bin/phd debug task --pool 4 now scales down cleanly instead of trying to
relaunch the second daemon after its scales down if the first one is
hibernating.

Auditors: chad