It depends on what you want to self host.

As an example, a family member self hosted home assistant. They didn’t have to know anything really. That was all they were doing and they bought the canned implementation.

If you have multiple services, you may need to know nginx configuration with virtual hosting.

You may want to use podman or docker or kubernetes.

It all depends …

Yeah, but it is hard to separate that, and it’s easy to get a bit resentful particularly when a projects quality declines in large part because they got lazy by duct taping in container registries instead of more carefully managing their project.

You’ve been downvoted, but I’ve seen a fair share of ZFS implementations confirm your assessment.

E.g. “Don’t use ZFS if you care about performance, especially on SSD” is a fairly common refrain in response to anyone asking about how to get the best performance out of their solution.

Actually, the lower level may likely be less efficient, due to being oblivious about the nature of the data.

For example, a traditional RAID1 mirror on creation immediately starts a rebuild across all the potential data capacity of the storage, without a single byte of actual data written. So you spend an entire drive wipe making “don’t care” bytes redundant.

Similarly, for snapshotting, it can only track dirty blocks. So you replace uninitialized data that means nothing with actual data, the snapshot layer is compelled to back up that unitiialized data, because it has no idea whether the blocks replaced were uninialized junk or real stuff.

There’s some mechanisms in theory and in practice to convey a bit of context to the block layer, but broadly speaking by virtue of being a mostly oblivious block level, you have to resort to the most naive and often inefficient approaches.

That said, block capacity is cheap, and doing things at the block level can be done in a ‘dumb’ way, which may be easier for an implementation to get right, versus a more clever approach with a bigger surface for mistakes.

Yep, and I see evidence of that over complication in some ‘getting started’ questions where people are asking about really convoluted design points and then people reinforcing that by doubling down or sometimes mentioning other weird exotic stuff, when they might be served by a checkbox in a ‘dumbed down’ self-hosting distribution on a single server, or maybe installing a package and just having it run, or maybe having to run a podman or docker command for some. But if they are struggling with complicated networking and scaling across a set of systems, then they are going way beyond what makes sense for a self host scenario.

Based on what I’ve seen, I’d also say a homelab is often needlessly complex compared to what I’d consider a sane approach to self hosting. You’ll throw all sorts of complexity to imitate the complexity of things you are asked to do professionally, that are either actually bad, but have hype/marketing, or may bring value, but only at scales beyond a household’s hosting needs and far simpler setups will suffice that are nearly 0 touch day to day.

For 90% of static site requirements, it scales fine. That entry point reverse proxy is faster at fetching content to serve via filesystem calls than it is at making an http call to another http service. For self hosting types of applications, that percentage guess to go 99.9%

If you are in a situation where serving the files through your reverse proxy directly does not scale, throwing more containers behind that proxy won’t help in the static content scenario. You’ll need to do something like a CDN, and those like to consume straight directory trees, not containers.

For dynamic backend, maybe. Mainly because you might screw up and your backend code needs to be isolated to mitigate security oopsies. Often it also is useful to manage dependencies, but that facet is less useful for golang where the resulting binary is pretty well self contained except maybe a little light usage of libc.

But it you already have an nginx or other web server otherwise required to start up (which is in all likelihood the case), you don’t need any more auto startup, the “reverse proxy” already started can just serve it. I would say that container orchestration versioning can be helpful in some scenarios, but a simple git repository for a static website is way more useful since it’s got the right tooling to annotate changes very specifically on demand.

That reverse proxy is ultimately also a static file server. There’s really no value in spinning up more web servers for a strictly static site.

Folks have gone overboard assuming docker or similar should wrap every little thing. It sometimes adds complexity without making anything simpler. It can simplify some scenarios, but adding a static site to a webserver is not a scenario that enjoys any benefit.

Because serving static files doesn’t really require any flexibility in web serving code.

If your setup has an nginx or similar as a reverse proxy entry point, you can just tell it to serve the directory. Why bother making an entire new chroot and proxy hop when you have absolutely zero requirements beyond what the reverse proxy already provides. Now if you don’t have that entry point, fine, but at least 99% of the time I see some web server as initial arbiter into services that would have all the capability to just serve the files.

Gitlab is an interesting project.

Admittedly looks good and has the feature set, but…

You dig in at all it’s a mess. If it screws up, it’s a mess. Every five minutes they seem to have another extremely severe security issue. Even when it’s working right, it’s a resource hog.

Has the feel of a lot of mismatched stuff duct taped together, but the duct tape looks great. So as a user, great. As an admin, a bit of a mess.