From stakeholder spreadsheet, requested by Samy:

Debugging while creating a new workload is difficult. It's hard to tell if the issue is related to the code for a new workload, a bug in DSI, or a bug in the server. For example, sometimes it's unclear why a perf workload failed and requires searching through all the logs carefully to find the issue. There are a lot of red herrings, so maybe one possible idea is to create a debugging guide.

Proposed solution: Make sure error logs explicitly mention where the error comes from, e.g., infrastructure versus user workload. Update docs if necessary to clarify any errors.

Timebox this investigation to 2 days.