There's a fairly wide history of self hosted Government Microsoft Exchange servers ending in ransomware payouts that suggests that blunt self hosting rules isn't as useful as it sounds.
And right now you look can look all the big MoveIT mass exfiltration hitting a wide range of orgs that wouldn't use Sharepoint "for security reasons".
In India, anybody can get a business account, which allows you to setup a catalogue of things, services, and you can pay inside app through underlying UPI interface (government operated quick guaranteed no-return money transfer). I have booked cars for rental, I have bought grocery, pizza, shoes n stuff.
WhatsApp was used by the Afghan Army to coordinate military communications and maneuvers. I wouldn't be surprised if other under-resourced military forces (i.e. Ukrainian) use it as a form of communications.
Afhanistan and the Ukraine both have a little less than 40M inhabitants. (Which is around half of germany, or like 20M less than France).
The Ukraine has/had a relevant IT-offshoring sector, from Wikipedia there's around 1/3 of the soviet military industry that was sitting there (including companies like Antonow)... and they had a few years of preparation since the krim was occupied where they knew that russian secret services would try to undermine any kind of communication infrastructure.
I doubt that they currently rely on WhatsApp for military communication...
dig +trace reveals that only some of the nameservers don't respond (on IPv4, but got a fine response after trying an IPv6).
;; communications error to 66.111.50.12#53: timed out
;; communications error to 66.111.50.12#53: timed out
;; communications error to 66.111.50.12#53: timed out
;; communications error to 66.111.48.12#53: timed out
whatsapp.com. 60 IN A 157.240.223.60
;; Received 57 bytes from 2620:13e:100d:c::35#53(d.ns.whatsapp.net) in 119 ms
Especially a.ns.whatsapp.net still seems down (IPv6 also).
I wonder whether this could be a BGP thing again, it was modified very recently according to [1] but I don't understand too much about that.
I didn't get the sense of a BGP error, mtr to ipv4 [a-d].ns.whatsapp.net looked ok enough during the outage. I even got back responses from d.ns.whatsapp.net to mtr probes while not getting DNS responses. Although, now that DNS is working, the mtrs are pretty different, I'm not used to getting such wide geographical response.
Disclaimer: I worked at WhatsApp until 2019. We built in resilience against DNS failing because there's a lot of networks in the world, and some of them don't have consistently working DNS. Turns out it also helps a lot when someone hijacks your ___domain account (grr), or DDoSes your DNS provider (on my birthday too, grr), or you mess up DNS somehow (nobody was willing to tell me what happened yesterday, grr). Doesn't help much with BGP based outages though.
The trick is if you bake IPs into your client packages, you need to have most or at least some of those IPs still working until the packages expire, and you need to have a process for server teams and client teams to update the lists for future releases as well as for server teams to know what's in the client builds in the wild. That's not too hard if you're running in bare metal hosting and your hosts are long lived. It's harder if you're running ephemeral cloud stuff (but there's ways to reserve IPs, so it's not _that_ hard).
There's some additional difficulty if you're only running ipv4 servers and some of your clients only have NAT64 access to ipv4. But, it's 2023, you can and probably should make ipv6 work on your servers.
If it's just DNS, as it seems to be from my perspective, that's likely to break the gateway for browsers, but it doesn't mean the gateway is down; just hard to access.
The phone client is resilient to DNS failures, but it's hard to do that for a website used by a browser (I'm not really sure how it would be done, probably something with service workers which can intercept and process outgoing requests?); and impossible to do if you hadn't loaded it previously.
Disclaimer: I worked at WhatsApp until 2019, including on resilience to DNS failures.
Yeah, it's possible to have IP fallbacks in the downloadable app, but if it's not working, it's likely it wasn't done (although I'm not in a place to check if the web gateway is currently working; it's possible two things are broken at the same time)
Looks like it's back now (or in progress to back anyway), but if you want to mess around with your hosts file, and you don't mind a bit of spelunking, you can find your suggested facebook IP with host www.facebook.com, and then look around in the /24 of that IP, looking for whatsapp-cdn-... in reverse DNS. That IP should respond well to requests for web.whatsapp.com. You might need to do something for the websocket hostname(s) too, but I forget the details of that. Might work going to the same IP. Take a look at domains used now, in case you want to do this again in the future.
It happened just as I was connecting my new phone to Whatsapp. Made for a few worrying moments where I was wondering if I had made a mistake somewhere and got myself locked out of my account...
Mostly Eastern European countries. Some fringe far right groups in Western Europe use Telegram because of a percieved higher level of privacy. If WhatsApp went down for any considerable length of time, countries like Spain would fall apart. I do recall an outage of several hours a couple years back that contributed to a huge spike in Telegram and Signal installs.
Then, sadly it went back up and everyone just went back to WhatsApp.
