Seems unrelated to their infrastructure, the DNS records for facebook.com, instagram.com, whatsapp.com and all derivative domains are wiped clean it seems
edit: though saying that, they do run their own registrar... Might've fucked something up over there.
WhatsApp and Instagram are both in FB infra. As I understand it, Instagram is fairly integrated with FB services; when I left in 2019, WhatsApp was less so, it was mostly WhatsApp specific containers running with FB's container orchestration on FB machines dedicated to WhatsApp (there was and probably is some dependence on FB systems for some parts of the app, for example the server side of multimedia is mostly a FB system with some small tweaks and specific settings, but chat should be relatively isolated). Inbound connection loadbalancing is shared though.
FWIW, WhatsApp (on phones) should be resiliant to a DNS only outage, the clients contain fallback IPs to use when DNS doesn't work, and internal services don't use DNS as far as I remember.
At one time, WhatsApp had actually separate infrastructure at SoftLayer (IBM Cloud now), but that hasn't been in place for quite some time now. When I left, it was mostly just HAProxy to catch older clients with SoftLayer IPs as their DNS fallback.
How could they all go down at the same time, if they have different teams of engineers running each product separately?
Could anyone with some background (or person familiar with the matter) explain how their system's set up?