emit metric for probe health

[philbrookes]

Currently there are no metrics (and kube state metrics are going away, see: Kuadrant/kuadrant-operator#675) emitted from the dns-operator to determine whether or not a probe is currently healthy, why it is failing or how many times it has failed.

We should add a metric which will allow us to determine how many healthy / unhealthy probes we have, and possibly include some degree of explanation (@david-martin has mentioned that a free string in a label is a bad idea, but we could aggregate the reasons into generic buckets.).