Add a common fault-tolerant mechanism of interaction with the elassandra

[YodaMike]

We have often used object mapper for elassandra without error handling, that was good at first, but when the service is experiencing a high load there are a lot of timeout exceptions, so we need to add a common mechanism for handling such errors. After timeout exception is cathed we need to give to elassandra some time to restore and remain our queries after a certain period.