Quoting Adam Borowski (kilobyte@???):

[the likes of GitLab:]

> This wouldn't have helped. A piece of infrastructure that's run by a person
> who gets run over the bus, or throws a tantrum, is inaccessible just the
> same as when hosted on GitHub.
>
> You need redundancy.

I'm puzzled: What about the likes of GitLab precludes redundancy? In
that particular instance, there are gobs of detailed documention devoted
to that subject, e.g.,
https://docs.gitlab.com/ee/administration/high_availability/

I take the crucial need for fallbacks and elimination of SPoFs so much for
granted that I seldom see a need to belabour that point.