We need to document all the possible metrics that we would like to expose for helping ing debugging out DNS service.
A few metrics that come to mind for `dcos-overlay` on each agent:
a) Total records for each DNS zone.
b) Total cummulative records.
c) Total DNS requests processed in some time period (second, minute, hour, day)?
d) Total DNS requests forwarded to `mesos-dns`.
e) Total DNS requests forwarded to upstream DNS resolvers.
f) Total DNS requests for each TLD (*.mesos, *.com, *.k8s)?
These are just a few.