[Feature]: Improve testing strategy #373
Description
Prerequisites
- I searched existing issues
Feature Summary
Currently the tests run for ~60m per PR which is not ideal. This means that every PR takes a really long time to be merged in. We should improve our testing strategy.
We should try and target <30m for end-to-end tests to finish.
Problem/Use Case
Better development experience
Proposed Solution
There are several options possible:
- split the test cases into core and full, run the core pre-merge and full post merge
- split the tests that don't depend on arch 50-50 and run them on arm and amd runners
- kick off a github parallel workflow to run x% of tests in parallel
- remove tests that can be tested in unit tests
- combine tests that are duplicate
- combination of the above
Component
Multiple Components