Dyn will still match them, even while they're empty before we've had a
chance to add the respons pools to them which is BAD and caused a
medium-size outage (thankfully not as bad as it could have been.) Ideally
we'd use publish=False and stage things, but that seems broken in the client
lib, there's no way to send publish=N. I also tried sending the
response_pool_ids as part of the create calls and response pool config if
one didn't exist, but neither of those routes worked :-(

* spelling: ancillary
* spelling: antarctica
* spelling: australia
* spelling: authentication
* spelling: continental
* spelling: constructor
* spelling: conversion
* spelling: creation
* spelling: doesn't
* spelling: easily
* spelling: efficiently
* spelling: equivalent
* spelling: essentially
* spelling: everything
* spelling: exactly
* spelling: be
* spelling: expensive
* spelling: supports
* spelling: healthcheck
* spelling: immediately
* spelling: ignored
* spelling: invocation
* spelling: itself
* spelling: leftovers
* spelling: missing
* spelling: natural
* spelling: nonexistent
* spelling: peculiarities
* spelling: pointing

This change hit a line length limitation, so I'm wrapping it and
adding a period which appears to match local style...

* spelling: quicker
* spelling: response
* spelling: requested
* spelling: redirect
* spelling: traffic
* spelling: unknown
* spelling: uploaded
* spelling: useful
* spelling: separately
* spelling: zone

Would have prevented both A and AAAA exisiting on the same node.

More thorough unit tests while I'm in here. Ended up doing some hacks/monkey
patching of DSFMonitor as the client library's object doesn't seem to be fully
functional/useful and has inconsitent behavior.

Also fixes some mocking data to match what the Dyn client libs are
expecting.