Because we're seeing random test failures, I wonder if you could temporarily enforce serial execution of all tests, just to rule out that parallel execution of tests might cause trouble.
Once we see all tests working (or tests failing despite serial execution), the workflows could be reverted to parallelism.
What do you think @ni4 ?
Because we're seeing random test failures, I wonder if you could temporarily enforce serial execution of all tests, just to rule out that parallel execution of tests might cause trouble.
Once we see all tests working (or tests failing despite serial execution), the workflows could be reverted to parallelism.
What do you think @ni4 ?