I share your generic concern for not using SQL keywords like GROUP as a name, and I think the Pi-hole development team does so as well - but I also think that point has been made quite early on in this discussion.
However, I do not share your take on the custom solution you are using. (click for details)
Pi-hole itself does not claim to support syncing multiple instances, and neither does it make any assertions towards supporting those.
At the same time, it does not forbid third parties to develop a syncing solution.
Pi-hole's source code is publically available, you can build on that.
stevejenkins/pihole-cloudsync dates about two years back.
Back then, Pi-hole 4 was using pihole-FTL.db only, and gravity was a term for a collection of blocklist files. Naturally, that solution would have relied on file based operations.
gravity.db was introduced along with client-based filtering and group management some 8 months ago with Pi-hole 5 in May last year, changing file based operation into database interactions.
That allowed for a major step forward in terms of fine-grained control of DNS blocking if you care to make use of it, while it will just function as usual for those users that don't.
That major shift didn't happen in the dark:
There has been an extensive public beta-phase starting in January 2020, four months before Pi-hole 5 was rolled out.
That would have been the right time for 3rd party tool maintainers to familiarise with the changes and raise their concerns, and let me assure you that the Pi-hole development is always more than keen to listen (though it may take strong, valid arguments to convince them of your ways).
It is still possible to sync two Pi-hole instances with Pi-hole 5.x.
If you are not depending on pihole-cloudsync, have a look at other community solutions. e.g. GitHub - vmstan/gravity-sync: š« The easy way to synchronize the DNS configuration of two Pi-hole 5.x instances..