I don't think rollback is something that can be considered for automation,rollback is sort of last resort type of thing, if things goes so wrong that it's not possible to fix stuff easily and quickly. It was and is automated, they got a primary database, and a "backup" database; in case the primary should ever crash/stop/whatever the backup will be picked up directly; That's why people could login like within 15 seconds; after the DC. So I believe currently the latest version (Monday morning) is now in "error" state and the latest database snapshot (seemingly from Saturday) is now "live". Clearly this was a automated process to keep the server up, and should probably not have happened. If they can undo the rollback depends on some things :1) what is the issue with the primary server (can it be resolved, and how fast)2) what is the latest backup (if not friday/saterday) I assume if they got a daily backup, that should be placed back, to have reduced impact; however if friday/saterday was the latest backup there is nowhere to go back to and all in game data changes are lost since then; We need to wait to see what devops can do, and what went wrong; based on that Anet is going to have to decide what to do; Beside, I think allot of engineers are having a horrible night/morning, adding the COVID issues I think this is a difficult time for them.