@jh@social.trom.tf @chjara@mk.absturztau.be
> How could that error occur suddenly when it did and how would it delete Facebook, Whatsapp, etc. DNS entries?

Just answering this part, because I have some.. experience.

I suspect something like this:
1. FB almost never uses this part of their infrastructure (since it needs changes only very rarely).
2. The infrastructure is shit and not standardized because of 1.
3. There was some stupid interpretation of `None` (or equivalent, e.g. `[]`) as an argument to a config call (e.g. "make this change in `[]`" meaning "everywhere" instead of "nowhere"; this is based on actual events in another big company).
4. Someone ran something like 3, perhaps automatically, and due to 2 it didn't do the thing they expected.

And bam, global idiotic looking outage without anything particularly mysterious.