Add checks to each scenario to ensure push server returns proper HTTP status codes

[rpappalax]

As-is, ap-loadtester can be used as a blunt-instrument to 'hammer' autopush in a variety of scenarios at scale. If the server falls over or throws a traceback, that's the closest thing to a "fail" that we can currently measure. The closest thing to a test "pass" is the absence of a traceback.

Ideally, (for API testing), we'd walk through the various scenarios and ensure that autopush returns us the HTTP status code(s) we expect with each step. Currently the only way this information could be accessed would be tracing a given test all the way back through Kibana.