While creating backup, the hypershift OADP plugin temporarily pauses reconciliation of HostedCluster, CAPI Cluster, NodePool objects. 

In the scenario where a backup fails, this can lead to leaked paused states. 

Example failures:
- backupstoragelocation secrets being removed
- any type of generic pod failures or workload interrupts (drains?) 

N/A     

    Unknown

    1. Create backup using hypershift oadp plugin
    2. interrupt / fail the backup e.g. by deleting the velero pod (untested) or by deleting the backupstoragelocation credentials (seen on production)
  
    

    Reconciliation of hypershift / control plane resources remains in a state with `paused: true` permanently. 

    Reconciliation of hypershift / control plane resources is back to `paused: false`, even if a backup fails. 

    Suggestion is to implement something like a post-hook that ensures any pauses triggered through backups are removed.