[Feature/Operator] Scaledown offloading choice

[rayyno]

Feature Request

Is your feature request related to a problem? Please describe.
The current operation to scaledown a node through the operator is disconnecting, offloading and deleting the node.
The offloading takes too long for large data clusters when removing some nodes from the cluster, causing the cluster to get stuck until offloading is done.

Describe the solution you'd like to see
An option to choose whether to offload a node when scaling down would be great.