The OS SSD died on my kubernetes/rook/ceph cluster which left the cluster with catastrophic monitor store loss. The process for recovering from just OSDs is not well documented, especially not when combined with rook and cephfs with multiple data pools. The last of which may never have been done before.
I am writing this document after the fact, so I may be missing some things, or they may be out of order. I will try my best though as I think it could be helpful.