Traditional server-based self-hosting will have lower average uptime, will be easier to attack, and will have a much higher chance of disappearing out of nowhere (bus factor event, or for any other reason).
A decentralized or distributed solution would make more sense as a suggestion here. Radicale (this one) is such an effort I’m aware of, although I never tried it myself or take a look at its architecture.
Traditional server-based self-hosting will have lower average uptime, will be easier to attack, and will have a much higher chance of disappearing out of nowhere (bus factor event, or for any other reason).
It’s not a single point of failure at least but if your particular project is targeted then yeah. I was thinking more about using it for private repos, where it isn’t public at all but that’s a separate case.
This is a good argument for self-hosting Forgejo (which is quite simple compared to gitlab from what I hear).
But good to see they are standing up to this shit.
Traditional server-based self-hosting will have lower average uptime, will be easier to attack, and will have a much higher chance of disappearing out of nowhere (bus factor event, or for any other reason).
A decentralized or distributed solution would make more sense as a suggestion here. Radicale (this one) is such an effort I’m aware of, although I never tried it myself or take a look at its architecture.
It’s not a single point of failure at least but if your particular project is targeted then yeah. I was thinking more about using it for private repos, where it isn’t public at all but that’s a separate case.
So much simpler than gitlab. An executable and a single config file. That’s all there is if you use sqlite as the database.
Gitlab was a farmyard of different things to worry about.