Don't know but I wouldn't bet on the "surely they have 100% redundant hot-standby systems", and if anything the last two years has been about bunkering down and avoiding costs rather than taking it as an opportunity to invest. I've no idea what they've got 'under the hood', but providing full site resiliency with mirrored hardware I'd suspect to cost 2.5-3x as much as a single site, both in direct cost and ongoing operational cost. And even that isn't a 100% guarantee against human error/bugs making it to production. If they didn't have that in 2019 I wouldn't imagine many would have pitched it between March 2020 and now.

Doesn't mean they shouldn't do it of course, and I'd really hope someone's at least tried running the numbers on it.

[edit] CWS beat me to it; geez... that's a little embarrassing if losing one server/array took them down[/edit]