One of the Biggest Crypto Exchanges Goes Dark and Users Are Getting Nervous
bloomberg.com
bloomberg.com
The only thing I can think of would be some big database migration that went awry and they don't have strategy to roll back.
But normally you would get a dump of production database and test your migration scripts on test/staging servers against latest snapshot of prod data.
You don't just run migration script on your production database and see what happens. You try running it against last backup / snapshot of prod data first.
What else could it be?
Either way, it looks bad from both engineering and product perspective.