Fixing connections

• maintaining one pool per node (rather than one pool for all nodes) helps in scenarios where one of the redis node goes down.