That depends on a lot of factors, but for me personally, yes it is. Much worse.
Assuming we’re talking about hosting things for Internet users. My fiber internet connection has gone down multiple times, though relatively quickly restored. My power has gone out several times in the last year, with one storm having it out for nearly 24 hrs. I was sleep when it went out and I didn’t start the generator until it was out for 3-4 hours already, far longer than my UPSes could hold up. I’ve had to do maintenance and updates both physical and software.
All of those things contribute to a downtime significantly higher than I see with my stuff running on Linode, Fly.io or AWS.
I run Proxmox and K3s at home and it makes things far more reliable, but it’s also extra overhead for me to maintain.
Most or all of those things could be mitigated at home, but at what cost?
I've been dabbling in this since 1998, It's almost always ISP and power outages that get you. There are ways to mitigate those (primary/secondary ISPs, UPSes, and generators) but typically unless you're in a business district area of a city, you'll just always be subject to problems
So for me, extremely anecdotally, I host a few fairly low-importance things on a home server (which is just an old desktop computer left sitting under a desk with Ubuntu slapped on it): A VPN (WireGuard), a few Discord bots, a Twitch bot + some auth stuff, and a few other services that I personally use.
These are the issues I've ran into that have caused downtime in the last few years:
- 1x power outage: if I had set up restart on power, probably would have been down for 30-60 minutes, ended up being a few hours (as I had to manually press the power button lol). Probably the longest non-self-inflicted issue.
- Twitch bot library issues: Just typical library bugs. Unrelated to self-hosting.
- IP changes: My IP actually barely ever changes, but I should set up DDNS. Fixable with self-hosting (but requires some amount of effort).
- Running out of disk space: Would be nice to be able to just increase it.
- Prooooooobably an internet outage or two, now that I think about it? Not enough that it's been a serious concern, though, as I can't think of a time that's actually happened. (Or I have a bad memory!)
I think that's actually about it. I rely fairly heavily on my VPN+personal cloud as all my notes, todos, etc are synced through it (Joplin + Nextcloud), so I do notice and pay a lot of attention to any downtime, but this is pretty much all that's ever happened. It's remarkable how stable software/hardware can be. I'm sure I'll eventually have some hardware failure (actually, I upgraded my CPU 1-2 years ago because it turns out the Ryzen 1700 I was using before has some kind of extremely-infrequent issue with Linux that was causing crashes a couple times a month), but it's really nice.
To be clear, though, for an actual business project, I don't think this would be a good idea, mainly due to concerns around residential vs commercial IPs, arbitrary IPs connecting to your local network, etc that I don't fully pay attention to.
Unanswerable question. Better to perform a failure mode analysis. That rack in your basement would need redundant power (two power companies or one power company and a diesel generator which typically won't be legal to have at your home), then redundant internet service (actually redundant - not the cable company vs the phone company that underneath use the same backhaul fiber).
