I know Postman has kinda gone to shit over the years but it's hilarious my local REST client that makes requests from my machine has AWS as a dependency .

I found that out about Plex during an outage too.

It's seemingly everything. SES was the first one that I noticed, but from what I can tell, all services are impacted.

In AWS, if you take out one of dynamo db, S3 or lambda you're going to be in a world of pain. Any architecture will likely use those somewhere including all the other services on top.

If in your own datacenter your storage service goes down, how much remains running

Agreed, but you can put EC2 on that list as well

Assuming running instances remain running, ec2 would be less bad I think.

yes, the previous AWS incident to this one iirc ec2 instances were not reachable from AWS console (IAM issue?) but they kept running and working

When these major issues come up, all they have is symptoms and not causes. Maybe not until the dynamo oncall comes on and says its down, then everyone knows at least the reason for their teams outage.

The scale here is so large they don't know the complete dependency tree until teams check-in on what is out or not, growing this list. Of course most of it is automated, but getting on 'Affected Services' is not.