We should parallelize our test suite as much as possible. This will allow for much faster test runs and uncover potential bugs in our implementations of concurrency.

Some existing tests may not be parallelizable, and fail points are still not build for concurrency, so this may require a larger refactor.