Unless something has changed recently, all you can do is set budget alerts on billing updates. Runaway costs for people simply testing AWS is common. (On the bright side, again unless something has changed recently, asking them in support to scrap them works.)
Both of these were problems for the author. I don't mean to "blame the victim" but the choice of AWS here had a predictable outcome. Static documentation is the easiest content to host and AWS is the most expensive way to host it.
Really high bandwidth costs in general. I've never worked anywhere large enough to hit them, but I've heard inter-AZ traffic in the same region can become quite expensive once you're big enough
This is true. There are services that force use of multi-AZ deployment, like their version of Kafka, or basically anything that creates autoscaling groups between AZs (like EKS). Without tight monitoring stuff can get out of hand fast.
What surprised me is you get charged both ways. $0.01/GB egress out of the source AZ and $0.01/GB ingress into the destination AZ. So it's easy to underestimate the billing impact by half.
True, those are also a foot gun--especially for AWS services where you'd want to, instead, switch to VPC Endpoints above a trivial amount of traffic (I think the endpoint interface is like $2/mon so has a small cost)
In addition, it's a pay-per-use platform