FEATURE: Operator rolling restart on disk resize

Currently, the Vitess Operator does not automatically perform a rolling restart of vttablets when the disk size of a keyspace is increased. As a result, we need to manually delete pods and carefully track which pod is the primary to ensure it is deleted last.

This manual operation becomes especially painful when dealing with multiple shards, turning it into a significant operational headache ("nightmare") that is error-prone and time-consuming.

Proposed Feature:
Introduce functionality in the Vitess Operator to automatically trigger a rolling restart of all vttablets in a keyspace after a disk resize. This would ensure that pods are restarted in the correct order, maintaining primary availability and reducing the risk of manual errors. Similar as it does when keyspace resources are increased or config updated.

Start by examining the existing rolling-restart behavior used when keyspace resources increase or configuration is updated. Trace how disk-size changes are detected and how vttablets are ordered, then verify that a disk resize triggers an automatic restart while preserving primary availability across multiple shards.