Currently, operator uses pod eviction during statefulSet rolling update. It makes hard to use podDisruptionBudget with it.

Disruption budget is useful for vmstorage operation and it instructs other Kubernetes components how many replicas should be kept.
But for minimal downtime it's crucial to remove all pods at once, which pod disruption will prevent.

Possible solution could be to switch for pod.Delete, if statefulset unavailable pods is equal to 100%.