HA trident-controller

[KreshnikZ]

Description: The trident-controller runs as a single replica deployment. Given the importance of the controller pod for all functions, including volume mounting, it would make sense to run two replicas of the controller so that during cluster upgrades or if one pod terminates for some reason, there is another running pod. This should be combined with a podDisruptionBudget.

Feature request: The trident-controller deployment should have 2 replicas. In addition a podDisruptionBudget object can make sure that there is one available replica.