How are AWS/GCP/Digital Ocean metadata services implemented?

I poked around some AWS EC2 and GCP compute instances. Noticed a 169.254.169.254 addresses. I was wondering how they implement their meta data service at scale.

There's some literature and some videos that explains how AWS VPC is implemented from a low level perspective. I was wondering if there's any other talk like that for AWS Metadata service.

I found literature explaining an open stack implementation of a metadata service. This picture basically sums up the explanation.

Is the way aws and other cloud providers do it similar? Some insight would be grateful.