1. Proposed title of this feature request

MachineSet: Enable "Migrate" for N2D instances instead of "Terminate" on GCP

2. What is the nature and description of the request?
Summary:
OpenShift currently requires the onHostMaintenance field in MachineSets to be set to "Terminate", which leads to VM reboots during GCP infrastructure maintenance. This causes application disruptions in production environments. However, N2D instances on GCP support VM migration, and a patch has been proposed to support "Migrate" for these instances.

(https://docs.redhat.com/en/documentation/openshift_container_platform/4.16/html/machine_management/managing-compute-machines-with-the-machine-api) and based on our experience, we must set MachineSet onHostMaintenance variable to "Terminate" and cannot configure it to "Migrate".

Though there is an exception for N2D instances on GCP (See documentation : https://cloud.google.com/confidential-computing/confidential-vm/docs/troubleshoot-live-migration).

Proposed a pull request on machine api operator to enable maintenance VM Migration for N2D instances : https://github.com/openshift/machine-api-operator/issues/1405
Problem must also be present on openshift-installer with first machines (bootstrap, masters) and first machinesets (based on install-config.yaml file).
A specific condition should also be included to check the CPU model (MinCpuPlatform = "AMD EPYC Milan").

3. Why does the customer need this? (List the business requirements here)

Application disruption during GCP infrastructure maintenance

4. List any affected packages or components.

Openshift, machine api operator