Overkill probably, but useful practise for when working on something of a scale where it will make a significant difference. And they might just like playing with these things. A home lab does not have to be DayJob practical!
Talking about overkill, he also use 4 physical machines each with a 4 core CPU. But its not about what you need, its about what you want to learn. So if you want to learn about cluster/orchestrating i think its warranted.
Some people just like to tinker and try other things even if it's overkill. I have a k8s cluster on DO to host shitty side projects, it's pretty nice to just `kubectl apply -f manifest.yml` and be done with it, especially since I'll probably wind up tearing everything I just set up anyway.
I fully agree that setting up Kubernetes is a lot of work, but I think people forget how much work it is to set up anything maintainable for more than a few services running on a handful of services.
The nice thing about Kubernetes is that it’s probably not much harder (if at all) than the alternative (assuming you already know which tools to use to compose your bespoke platform) and you can easily find online resources and support (never mind all of the problems you don’t run into because they’re already solved in Kubernetes).
I would rather choose something like xgp-ng as a hypervisor for a home lab environment such a this, instead of a custom one-off kubernetes environment.
Or just a total DIY xen or KVM hypervisor on debian or centos approach. But with considerably more RAM in the host machine than 16GB (like 64-128GB).
the kubernetes sort of makes sense if the purpose is also for the person who owns/operates it to test kubernetes things, learn kubernetes, or use it as a home lab of some larger similar environment that they have elsewhere.