Users need the ability to handle scenarios where Redis goes down, causing the internal redis-client actor to die.
Recipient[Actor[akka://TestActorSystem/user/redis-client-0#-1167390260]] had already been terminated.
I've added a test that demonstrates the failure.