A test-machine executes sequences of test commands

Test machines can be constructed to operate against a variety of targets. For
example:

  - a local development kafka cluster
  - a mock topology processor
  - a cluster shared with other users

In each of these cases, as a test-author we typically want to do the same type
of thing. Inject a bunch of events into the system, wait until the system under
test has finished processing, see what comes out the other end, and check that
looks good.

But the mechanism by which data is injected and observed is different in each
case. The test-machine exists so that the author doesn't care. The exact same
test (or scenario) can be executed against a mock topology processor, a kafka
cluster running on localhost, or (via smokin or the rest-proxy) a remote
kafka cluster shared with other users.

A test machine executor can be built by composing execution wrappers
defined in here