We need durability and reliability for work queues

WorkQueue usage failure points:
a.) Redis may crash
b.) Worker task or other consumer of queue may crash
resulting in lost work requests, especially if already removed from queue by the consumer.

With respect to issue a., see http://redis.io/topics/persistence.
With respect to issue b. see pattern: Reliable Queue in: http://redis.io/commands/rpoplpush.
