Investigation: can we further minimize the tests we run by default on releases PRs?

[yulianedyalkova]

Adding yet another set of upgrade tests made us question if we need to always run all the tests. As we also introduced release stages now, can we somehow optimize around that so we run a minimum amount of tests but still ensuring full coverage of the release?

Things to consider:

• We definitely want to do some detailed initial + final release validation
• Can we minimize the tests we run on the regular bumps?