When connector is used in the container environment (DC/OS, K8S, etc.), the IP of the nodes may change as they are restarted, or replaced. Even if we specify the contact points as host names, they are eagerly resolved to IPs at start, and this may cause problem later when SCC is trying to establish connection again, and old IP doesn't exist.
The Java driver 4.0 provides the possibility to delay resolving of the contact points until they are actually used - in this case, if Cassandra node that is specified as contact point changes IP address, it will be resolved anew on reconnect.
it would be useful to make this behavior configurable in SCC, as by default Java driver 4.x eagerly resolve host names to IP.