Because we take backups per cluster, if we need to restore QA, there's no straightforward way (to my knowledge) to roll just QA back without also rolling production back. Regardless, we should consider moving them to separate clusters so they are completely isolated. Minimally, we should make updates to the backup strategy (if possible) so that just QA or prod can be restored or document a method to restore QA or production separately.