systemd-nspawn / machinectl for OS containers, anyone using it?

Interested to know if anyone else out there is using systemd-nspawn / machinectl to manage full OS container images?

I have been investigating this method for a customer due to its less invasive nature (ie. minimal impact to base OS config.. I find this lwn post summarises it well).. and it is proving a good fit for full OS images rather than microservice / single process container workloads we use docker for.

It does appear to be a second class citizen to Red Hat's Docker aspirations, but still keen to hear if anyone using it at scale especially if you're using it as a migration path (step) away from traditional fully virtualised workloads? or even if you're using it in CI pipelines / build services etc.

I've run into a minor problem with SELinux in 7.3 and configuring machine startup through systemd, but fairly certain this is resolved in the 7.4 beta.