Add topic replication factor change

[fvaleri]

This change implements the proposal about topic replication factor change.
https://github.com/strimzi/proposals/blob/main/066-topic-replication-factor-change.md

Documentation will be added in a separate PR.

To avoid making the review of new components harder, there is a small amount of duplication related to Cruise Control client. This will be addressed in a follow up PR, where I plan to move the shared part between Rebalance and Topic operator in the operator-common module, and to switch from Vertx HTTP client to Java HTTP client.

Unit tests are included, but I also performed a number of manual tests on Minikube and OpenShift that could help in testing the feature, and creating system tests:

• Deleting secret $CLUSTER_NAME-cruise-control-api should roll CC
• Deleting secret $CLUSTER_NAME-entity-topic-operator-cc-api should roll both EO and CC
• Setting more replicas than available brokers should report an error in .status.replicasChange.message and reverting the change should clear that status
• Setting .spec.config."min.insync.replicas" < .spec.replicas should work logging a warning
• Bulk replicas change should take a reasonable amount of time (tested with 100 topics with 100MB each, and it took about 5-8 minutes on my local test environment)
• An ongoing replicas change should be recovered after TO crash
• An ongoing replicas change should be recovered after CC crash

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).