The insert approach creates exactly the test data a scenario needs, on demand. It works when automation can freely manipulate the service’s state; shared services or restricted APIs may prevent it. section 18.1 section 18.7
Scenarios can add custom entities to a clean-start or seeded baseline, using hooks or explicit context steps. This reduces dependence on entities already present in the baseline, but does not by itself guarantee isolation between parallel scenarios. section 18.1 section 18.7
Figure 18.1 Conflicting modifications when running tests in parallel using a shared database

Illustrates how modifications can conflict when tests run in parallel against a shared database. section 18.1
