We currently run PR tests in isolation i.e. when someone creates a PR, the PR tests are run against their changes. Once the tests are suiccessful, the code is merged into main branch.

This allows us to test code in isolation, but it means we have to queue PR’s up one after the other until the tests complete.

We could add another PR agent and allow PR’s to run in parrallel, but that could lead to code not being tested together. i.e. if 2 or 3 PR tests are running at once, it would only be testing that PR’s code changes and not code that could be merged in by the other PR’s.

So, my question is what is best practice in this situation and how can we go about running tests in parrallel to lessen the queue time waiting for testing?