Add ability to backup existing workload clusters in case of unrecoverable management cluster failures

Currently if the management cluster were to fail during upgrade, we store a backup from clusterctl for the management resources, but not the workload cluster objects from the eks anywhere api version. This makes it difficult to restore the workload cluster objects on a newer management cluster if we need to build a new one.

We should save the eks anywhere specific resources as a backup as well in the case that we lose our management cluster completely, otherwise it is difficult to try to recover those objects and have them point to the new management cluster. We can also explore having an import cluster option where we are able to generate the eks anywhere spec from a Kubernetes cluster based on information on that cluster, but that most likely will be a longer term task.

Start by tracing how clusterctl currently stores management-resource backups, then identify which EKS Anywhere API resources must also be preserved for workload-cluster recovery. Treat the import-cluster idea as a separate longer-term investigation; done should mean the required workload objects can be backed up and restored against a replacement management cluster.

Written by the indexing model from the issue text.