We are now returning real status codes from both services but they don't seem aligned, i.e. I've seen hundreds of 429's coming from the Evaluator and I'd expect to see this in the Orchestrator logs in the very least but were not there.

Additionally, we need to decide on how we officially (at least for now) want the Orchestrator to be responding to Evaluator errors based on the status codes, not just error messages as previously. We can devise a plan as a wrap-up to this task and create a followup afterwards.