Discuss and collect scenarios when it's taken a long time to do manual QA of an MR
From both my own observations and some discussions with the backend engineers in @gitlab-com/runner-group I've got a hypothesis that some of our challenges with merging things quicker stem from how long it takes to do some manual QA on the MR under review.
What I'd like to do with this issue is start collecting examples of when it's happened that the setup to test an MR took much longer than it felt like it should for an MR of that size.
This is ONLY about the manual QA challenge or time. Not about the static review of the code itself.
Example
Doing review for !123456 - to test it I had to compile from source "StevePods - the friendly alternative to kubernetes" because no binary was provided for my development environment. Once compiled and running, each test job took me 90-minutes to prep and configure. Overall this took me about 15-hours to test the basic functionality.