A question regarding our ability to better identify the point an upgrade failed in our e2e upgrade testing was raised and if we could collect clearer signal.

We need to understand what signal we get currently when an upgrade fails can we readily see the point of failure?

What can we do to get better signal on the failing steps?

Create specfic tests for failed steps in the upgrade process. 

Monitor CVO as the upgrade progresses to detect failure points

Leverage existing resource watch functionality

Creating spike story for further investigation