The solution around this is 2 fold:
Either use a front end load balancer that sends to the swarm nodes OR Configure the swarm nodes with like KeepaliveD to have a virtual IP address, and then point your DNS to that.
I'm personally using the second one for the past 4 years very happily.
When I'm interviewing people for any entry level, or even low level sysadmin positions, discussion of a homelab is always on my list of questions. I'm looking for not only the fact it exists, but also the passion you have talking about it. I can work on tech skills, but I can't teach passion.
They are very very bare bones .
Go look at Os ticket awesome theme instead
Low maintenance and self-hosted don't really go together. The reason those large companies are so successful in this space is that they are obfuscating and removing all of that maintenance, HA design, etc, which you're now responsible for when you self host.
What you can look into is designing the self-hosted solutions to be as low maintenance as possible. Design services in clusters, compared to single instances, to allow you to drop some for updates without impact. Design for failover. Design for proper backups. Design with application stacks in mind that are supported and widely used vs someone's random github you found. Design for modular setups (not just one giant Mariadb, multiple smaller ones, etc)
All of this is work, however, so it's a trade off on your side how much and where you're wanting to put the work in.