man, I’m really considering moving from us-east-1 to us-east-2. UE1 is a mess lately 😕 did anyone else have S3 issues today?

Most of us regard us-east-1 as the beta region. (All new stuff lands there first.) I only use it for the few things that must be managed there, like CloudFront.

CloudFront doesn’t work in other regions?

It's managed through us-east-1, and then distributed from there. (Last I knew, S3 was as well.) Lambda@Edge is more obviously only deployable to us-east-1.

hmm. most things seem fairly straightforward to move over, except Cognito

In other words, "global" = us-east-1 for managing, but then it operates all over the place. That's my impression anyhow (again transparent with Lambda@Edge), I don't know that it's documented anywhere.

yeah, that sounds right

I would recommend to not use us-east-1 is the oldest and messiest and biggest region of AWS

I wonder how easy it would be for me to seamlessly move my whole stack to USE2

thanks to sst and CDK maybe easier 😁

yeah. we can have VERY minimal downtime, like maybe 30 mins or so

well it really depends on the complexity of your architecture

Cognito is the toughest one, probably. not really sure if users can just be moved over to another region

maybe on the second you can play with the CDN

that’s true.. actually the API Gateway custom domains don’t really mean anything besides telling the API endpoint what hostnames are allowed to make requests. Route 53 does the rest of the work. so I can just point that wherever

I’ll have to talk to AWS support about Cognito

Sounds like painful 😞

yeah I saw that article before 😕 ugh

You can move easy stuff like Lambdas and DynamoDB but leave Cognito. It reduces your risk at least. Custom domain names can exceed your 30 minute downtime though. You can change it over quickly, but it can take hours to propagate to the world's DNS network.