Consider alternatives for cluster management based on cassandra

[AndrewX192]

Arcus Platform presently uses cassandra to solve agreement for clustering and potentially failover scenarios. Unfortunately this causes write losses with cassandra lightweight timeouts, and potentially signifies that cassandra is the wrong tool to use for solving this problem. Arcus already utilizes zookeeper for coordination of kafka, so this may be a more logical choice for solving quroum and consistency of multiple instances within a cluster, especially as failover support is implemented.

For now, this will be explored as an alternative implementation of the ClusterServiceDao interface, with an option to switch implementations through di.