Description
When using observe commands and a node is failed over and not rebalanced automatically, the code will time out because no broadcast will ever return but it expects one.
Proposed solution is to throw an exception to fail fast when there is no replica available.