cluster-api-provider-aws: Cluster with reused name in different namespace fails to boot

/kind bug

What steps did you take and what happened:

What did you expect to happen: A new VPC, set of machines, and associated resources should have been booted

Anything else you would like to add:

Environment:

About this issue

Currently we treat the cluster name as a unique value. We have a few options here:

Add some type of validation that the cluster name is indeed unique
- This would still present issues where the controllers are scoped to a single namespace or running under different management clusters
Prefix everywhere we use the cluster name for naming, tagging, etc with the namespace name
- This would still present issues where multiple management clusters are in use
- This could also lead to exceeding string lengths in various places and would need to be accounted for.
Generate some type of unique identifier that can be pre/post pended to the cluster name
- This unique identifier would need to be generated and saved to the spec to persist across pivot or a backup/restore (resource UUID does not persist, nor does Status)
- string lengths should also be accounted for
- may need to also tag resources with the namespace to help for easier identification of resources in AWS, or some other method to easily differentiate between clusters if running on a single management cluster.