Restore from backup fails to bring up healthy clusters for longer cluster and machine pool names

[ev-hines]

Rancher Server Setup

• Rancher version: 2.9.2
• Installation option (Docker install/Helm Chart): Helm Chart
• Kubernetes Version and Engine: 1.30.5+rke2

Describe the bug
Due to name length limits in kubernetes resources, the following resourceNameRegexp rule fails to find all possible machine-states. This causes restores to break for clusters due to empty secrets. Other resources in any regexp rules that search for a resource ending with a specific value are also likely to be affected by this issue.

    resourceNameRegexp: "machine-plan$|rke-state$|machine-state$|machine-driver-secret$|machine-provision$|^harvesterconfig|^registryconfig-auth"

To Reproduce
Steps to reproduce the behavior:

1. Create a cluster with a machine pool with names such that the sum of their lengths is >= 39 (a lower number of around 30-35 will likely work but this what I tested)
2. Create a backup
3. Tear down the rancher cluster and spin up a new one
4. Restore the cluster from the backup
5. The downstream cluster will never become healthy after restore and instead show a waiting for machine <project>/<machine_name> driver config to be saved forever.
6. The rke.cattle.io/machine-state for that node will show 0 keys
7. Check the backup tar ball and you will see there are no machine-state secrets in the backup.

Expected behavior
The secrets are properly backed up

Screenshots
Screenshot attached

Console output of secrets with 0 keys

fleet-default                                             osc-nonprd-test-test-control-plane-az1-jnt7g-ckwkv-machin-1be8a         rke.cattle.io/machine-state                   0      13m 
fleet-default                                             osc-nonprd-test-test-control-plane-az2-2sl6m-nv9h2-machin-d565f         rke.cattle.io/machine-state                   0      13m 
fleet-default                                             osc-nonprd-test-test-control-plane-az3-hmztw-dk65z-machin-c8097         rke.cattle.io/machine-state                   0      13m 

Additional context
As shown above, this seems to occur because the backup searchss for secrets to backup by name suffixes, not a label or secret type, which tend to get truncated when the name gets too long but still within allowed limits.