Investigate Test Flakiness

When rewriting the test framework I noticed two interesting tests that had problems (see below). Should investigate what might be causing the issue/flakiness.

• ClassicBucketManagementTests - test_cluster_sees_bucket
  • For whatever reason this can cause a segmentation fault, it is inconsistent and doing a basic script that opens multiple buckets via cluster.bucket() does not seem to trigger the seg fault. Need to investigate further, I suspect it has something to do w/ how the client is handling multiple open buckets.
• ClassicSearchTests - test_cluster_search_ryow
  • This test is flaky in that ~10% of the time it will fail, but pass on a re-run. From what I can tell the mutation tokens are passed properly to the C++ client, but should further investigate (i.e. verify tokens in the HTTP request?) to determine flakiness and make sure it is not an issue w/ the SDK(s).