I will merge it to master and schedule it for v1.54.1
There was, literally no difference between the v1.53 and v1.54 debug - it was just using a different TCP connection which isn't something which shows up in the dump bodies logs...
I think the resolution to this might be something Wasabi is interested in. I think it shows a potential weakness in their systems.
"I just wanted to give you an update and let you know I have opened up a defect on out end with the DevOps team to look into this. It's possible there is a configuration difference between the two regions (us-east-1 and us-east-2) which is why the rclone dev was unable to reproduce these errors in us-east-1. It may take some time to get the details figured out, but the data you and rclone provided are great reference points, and I appreciate the time and effort. I'll let you know if we discover any issues that need to be remedied, and if so, I'll reach out to you when we get them sorted out in case you wanted to run some more testing on your side."