This outage would've required cross region failover to be immune to. We'll see if customers switch to whatever company was resilient, but this has happened before and the answer was no.
The company I work for has a ton of stuff in us-east-1, many large products and sites, and we didn't go down. Our products/services aren't multi-region or multi-cloud. We don't pay exorbitant bills or have super complicated architectures.
That's the thing - most AWS services didn't "go down", as in stop working entirely. There were specific operations of specific services that were failing. Increased API error rates, inability to start new EC2 instances, billing metrics unavailable, AWS console unavailable, etc.
The outage wasn't like "all our servers stopped running". It was dynamic, new, specific operations that failed. If you just had a Fargate container that was started a week ago, and you have no need to restart the container today, it just kept chugging along.
Our architecture is stuff that just keeps chugging along. Fargate, S3, RDS, CloudFront, CloudFlare, etc. From our perspective, there was no outage in us-east-1. Literally the only alert we got the entire time was "billing limit exceeded" - and that was a false alarm, because it was set to alarm if there is zero billing data.
But is this strategy or luck? I'm not seeing how those many companies did something dumb or wrong here while you did it right. Like are they only affected because they overcomplicated their deployments? Either way, your service isn't resilient against a generalized regional outage it sounds like.
