Research and potentially Swap dnsmasq with coredns

[akrzos]

We need to replace dnsmasq on the bastion with coredns to see if the ANY query issue occurs. It plagues ipv4 and ipv6 disconnected clusters during the post deploy phase or anytime configuring alertmanager or anything else which depends on resolving an address that is not a kube endpoint while attempting to resolve it. Those queries get forwarded to dnsmasq and can bog dns down to a crawl causing many issues (that eventually resolve themselves)

[akrzos]

#342 is the first attempt to close this

[akrzos]

At this point it is safe to say that the default should just be set to coredns, and after several weeks of running that default it would be safe to remove the dnsmasq roles.