In order to achieve multi Availability Zone resilience, some customers create multiple ROSA Machinepools for a given Hosted Cluster, to have multi availability zone resiliance. The pattern consists in:

• One Machinepool per availability zone
• Single node per availabiity zone

If the customer does not have pod disruption budgets and maxSurge so additional nodes can be created, when they perform a MachineConfig change that requires Node replacement, the workloads running on those nodes can be affected.

This ticket is about thinking up if there's a need for multiAZ machinepools or any other pattern that could target this use case in a more workload preserving way.

Additionally, we might want to consider some sort of linting in status conditions for Nodepools if we see that Pod Disruption Budgets and maxSurges are not defined to warn the user about the repercussions of it while giving them the freedom of having things that way for their legitimate purposes.