My self-hosted git repo inside my own network is still working. How about yours?
@vkc Since no one else on this thread is representing this viewpoint: I want to remind people that unless you're putting unreasonable amounts of time, money, and energy into redundancy, self-hosted services have lower uptime than centralized ones. (At least, responsibly-run ones like GitHub.)
Sure, you maintain control. (Which IS valuable!)
But there's a cost in time and headaches.
@nephariuz "self-hosted services have lower uptime than centralized ones"
1. Citation needed, and
2. Uptime is only as important as the access need. I shut down servers when I don't need the services, another joy of self-hosting.
@vkc I'm betting in the past you've had issues (e.g. hardware failure) that you had to troubleshoot before you could access your own service. I certainly have.
For centralized services, such issues are *their* problem to solve. If you're self-hosting, it's *your* problem to solve.
@nephariuz yeah, but that's not what you said. What you said is that you "wanted to remind people" that the time spent on self-hosting your services well was "unreasonable". And what I'm saying is that your assertion lacks citation, plus ignores the fact that uptime isn't all that important to a lot of us.
@vkc Broad data is not available because people who self-host are not required to report uptime. And "unreasonable" is subjective, of course.
@nephariuz right, but you didn't say "subjectively there's an argument to be made".
You got on my post and felt like you "wanted to remind people", which *absolutely* comes across as arguing down to me, a literal professional, about the benefits of hosting your wares.
And if that's not how you intended it, please rethink your word choices and hop off this thread.
@nephariuz@mastodon.social The thing is that I'm not entirely convinced of that. I run a number of services for (literally) friends and family. Yes, I have invested time in the design and research of several things.
Mainly because I enjoy it, but that is the perception on my side.
For my friends and family as users, I wouldn't be absolutely sure that they experience more downtime than what they experience with centralized software.
As I don't need to run services for hundreds of users on every geographical area and given that a 99% uptime means if calculated through the whole year 14 daily minutes of downtime (https://uptime.is/), I would need to measure it but I would be surprised if my server is unavailable for three days and a half a year.