It depends on how the pipeline is setup. I have pipelines that are failing and some that are not. I have to adjust the ones that are failing to accept the non-zero return code from the scan stage in some instances. Thanks for creating the case.
My company is having this issue as well. The odd thing is that it is not a consistent issue on all of our PRs. It is only occurring on a few, but it is a pain to deal with. I also appreciate you working on this, but would like to see it resolved as soon as possible.