Comment by randomtask

So, they suggest that you spin up the dependent services as part of the test runner and run tests for your service against those? I've worked on a codebase that did this extensively. Over time it lead to a test suite that takes a long time to run, slowing down the release process considerably, and inevitably requiring further engineering effort to make the tests run in a reasonable amount of time. I can't say I would recommend it as an approach based on my experience with it, but YMMV.