Add documentation for workload cluster rollingupgrades side effect

**What would you like to be added**:
Currently, there are times where the upgrade of a management cluster to a new EKS-A release can have a side effect of performing a rolling upgrade on all the workload clusters managed by that management cluster. This side effect can be surprising and unpleasant for users to discover, and should be clearly stated upfront, with the research performed to verify the safety of those rolling upgrades. Although the EKS-A team does everything possible to prevent this side effect from occurring, there are certain configuration where it may be unavoidable.

**Why is this needed**:
Customers should plan their management cluster upgrades carefully to incorporate the possibility of this rolling upgrade

Review the existing EKS-A documentation for management cluster upgrades and workload cluster rolling upgrades, then research when upgrading a management cluster can trigger the side effect and whether those upgrades are safe. Done means users are warned before planning an upgrade, the relevant configurations and unavoidable cases are explained, and the safety findings are documented.

Written by the indexing model from the issue text.