It's very common to have to restart pixy and cache1 because both of them die unless kafka is running already. In addition the topic1-create job is a prerequisite of cache1 start.

Also during test runs we seem to have issues with requests back to the test's Express server taking too long.