Way too many unknowns at this point, it's entirely possible that it can be kept running on the basis of how well automated its infrastructure is. Which I'd "hope" it's very resilient, based on how well touted tools they're probably using like kubernetes are, but I've also never worked in Silicon Valley.
I've worked on information systems that did computer<->computer communication at a way larger scale than Twitter, it was a way more critical type of infrastructure and if I had to guess it's both more and less fragile than Twitter's infrastructure but probably in pretty different ways...
There's absolutely ways they could break the whole thing and cause outages or degrade performance/functionality for extended periods of time though, and with lots of sudden changes we're hearing about that seem pretty well-sourced, it could be days away. It could also just never really happen. But if they're changing up tons of teams and possibly permissions to the source control and management systems, hey, it's entirely possible someone hits the wrong button and blows up production because they thought there had historically been other approvers or something. Who knows ¯\_(ツ)_/¯
Yeah, it's all going to depend on how many safety checks are in the current orchestration and unit testing, and how well they're respected/documented. And whether people decide to override them just to shove something out the door.
I'll pray for a subtle bug that causes a slow degradation that eventually reaches a tipping point but has been going on so long that it's affected many pieces of production for as far back as they've got backups and can't just be dropped for some reason. I'm not sure if Twitter has customized their data storage layer at all though, like even if they're using a bunch of document databases or postgres and suddenly things are just in such a state that they're getting some obscure segfaults or something, it's still a system that there are a lot of people with expertise with who can probably figure out how to fix it.
Way too many unknowns at this point, it's entirely possible that it can be kept running on the basis of how well automated its infrastructure is. Which I'd "hope" it's very resilient, based on how well touted tools they're probably using like kubernetes are, but I've also never worked in Silicon Valley.
I've worked on information systems that did computer<->computer communication at a way larger scale than Twitter, it was a way more critical type of infrastructure and if I had to guess it's both more and less fragile than Twitter's infrastructure but probably in pretty different ways...
There's absolutely ways they could break the whole thing and cause outages or degrade performance/functionality for extended periods of time though, and with lots of sudden changes we're hearing about that seem pretty well-sourced, it could be days away. It could also just never really happen. But if they're changing up tons of teams and possibly permissions to the source control and management systems, hey, it's entirely possible someone hits the wrong button and blows up production because they thought there had historically been other approvers or something. Who knows ¯\_(ツ)_/¯
deleted by creator
Yeah, it's all going to depend on how many safety checks are in the current orchestration and unit testing, and how well they're respected/documented. And whether people decide to override them just to shove something out the door.
I'll pray for a subtle bug that causes a slow degradation that eventually reaches a tipping point but has been going on so long that it's affected many pieces of production for as far back as they've got backups and can't just be dropped for some reason. I'm not sure if Twitter has customized their data storage layer at all though, like even if they're using a bunch of document databases or postgres and suddenly things are just in such a state that they're getting some obscure segfaults or something, it's still a system that there are a lot of people with expertise with who can probably figure out how to fix it.
There's also always wildcards like the possibility someone hacking in specifically to break as many things as possible. Honestly it would probably hurt Tesla's share price by association so there's financial motivation for sure. Last time they got infiltrated, the person clearly had no real plan beforehand and just squandered it.