BrowserStack tests don't run correctly in case of multiple simultanous runs

Since we switched from TravisCI to github actions, the BrowserStack tests often give issues. It looks like when you push two branches at the same time, the tests for the two different branches are mixed up when running. Often, the BrowserStack tests hang and finally time out.

You can see for example that tests are mixed up from the number of tests completed/skipped, which is much too high (should be 6 times the actual number of tests because we're running on 6 different browsers):

Finished in 10 mins 26.368 secs / 9 mins 15.34 secs @ 12:22:21 GMT+0200 (Central European Summer Time)

SUMMARY:
√ 159765 tests completed
i 665 tests skipped
× 45 tests failed

Maybe we have to configure a unique id or something, I really have no clue.

Anyone able to look into this?