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.
Fixes an issue where we'd be looking for custom healthcheck config on the
existing record object (from the provider) which would never have a custom
setup. Instead looking at desired lets us find what's actually configured to be
the case
- This reworks the CallerReference structure for Route53 health checks in a
non-backwards compatible way. This means records will create new healthchecks
for themselves which should be fine.
- Since we're pre 1.0, support has NOT been added to cleanup the old
healthchecks. That could be done reasonably easy, BUT we'd have to keep that
around forever. The hope is that the new ref format/usage will prevent this
problem going forward since enough info exists in the ref to identify things
fully. :fingers_crossed:
- healthcheck GC is much cleaner and more robust thanks to ^
- overall the healthcheck management code is a bit easier to follow and more
robust now.
They were working around the lack of class hierarchy, this addresses that by
adding 2 child classes. It gets rid of a bunch of ugly default params and
if-this junk in the main class that was trying to deal with plain & geo records.
Also as a positive side effect it gets rid of some hacks that lived in
Route53Provider dealing with the fact that the old setup couldn't detect going
to/from a GeoDNS record correctly.
