When running an integration test, we have a constantly-running cluster. The egress connection establishment rate limiter tests assumes that stats will be zero for every test case, but the reality is that stats may increase in-between test cases running on a continuous cluster.

The change is to capture a baseline at the beginning of each test case to compare against.