User too easily can fill up queues

[cwjohan]

When queues get too large the redis-server may start rejecting new data or run out of memory.

Solution: Forward the 'drain' event from node-redis. The user can use it to know when to resume writing to a queue after a write has failed.

[cwjohan]

Fixed in version 0.2.2. The 'drain' event now is propagated. The user must otherwise manage flow control themselves.