A build of a patched `systemd-nspawn` and I've managed to get services running again.
Might just have to play about with Kubernetes on a separate host for the time being.