Could this be some sort of data corruption? I find it hard to believe that Slack could be down for this long without something that is exceedingly hard to rollback. Even if some services are completely overwhelmed with traffic, they could block a certain percentage of traffic to decrease load, and then force servers up across their datacenters and then unblock traffic. It has the hallmarks to me of some sort of datastore is down, but obviously just a random guess.