Hyper-V image fails to autoupdate to latest ... timeouts ... simply not robust

I was starting a new VM with the last Hyper-V image (clear-35000-azure-hyperv.*) but it fails to update.
The image starts in “autoupdate” mode and then never succeeds in updating. Log is full of errors “communicating with server timed out”. I happen to have run into this before and from what I read the solution was to try to upgrade in smaller increments. So that’s what I’m attempting to do. I was able to go to 35100 and I’m continuing from there. I was unable to go to 35200. Error was “Failed to retrieve 35200 MoM manifest”. It isn’t clear how to even find a small jump version to try. I’m trying 35300, but see timeout errors. Is there something wrong with the servers?

The other problem is that I needed to disable autoupdate before I could do this. You can’t have an autoupdate that continuously fails. Can Clear Linux figure out a robust solution for both “autoupdate” and these timeouts?

EDIT: there is a GitHub issue related to this which I added questions/comments.
swupd update behaves erratically (curl http/2 - related?) #2468