My issue with this 'yet-another-server' is that all of this stuff is new and doesn't have classic production qualities, or even have a version contract yet. Even OneDev is good, but I still staple the container version because sometimes he release incompatible minor or patch versions that are within contract. So you can't automate these things. This is why large companies have so much manual IT staff. I know so many people in this field now, literally manually installing certificates and even Windows updates from rdp. IT sux and nearly impossible to automate without risk of massive outages.

AlbyHub is still very experimental. I have no problem running an apg-get upgrade after a manual audit because nearly everything managed by Canonicle will usually be very stable, and the tools we might use from their prod repo will abide by production quality version contract. Meaning nearly no risk of massive downtime or corruption.

So basically every server we have to deploy is experimental, which means manual backups before and after, continuous babysitting for crashes, that can't be fixed and needs an issue to be created with the maintainers, manual version checking, package stapling, and monitoring.