Chris O'Haver

> I do not, however, think a completely flat structure will work well for us. Indeed, even the the example projects cited in the description (Envoy and Prometheus), do not...

I have updated the title of this PR to be more specific to what the changes are proposing. I hope this is OK!

> For example, from what I can see, Envoy has a ["senior maintainer" concept](https://github.com/envoyproxy/envoy/blob/main/GOVERNANCE.md#conflict-resolution-and-voting) where each "senior maintainer" have 2 votes while others have 1 vote. So this might also...

I think it's fair to say that in essence, this change is replacing final decisions by leader with final decision by maintainer vote. IMO, it's best for the CoreDNS community...

-1

I think that `Typify` is used both by cache, and in other contexts has been occasionally problematic, due to different expected behavior in various contexts. Not sure of the history...

For those following, this issue has been cross-posted to https://github.com/kubernetes/dns/issues/538 (where the node local dns cache version of CoreDNS is maintained).

> Sometimes i see strange error in CoreDNS logs How frequently does it happen? Can you estimate as a ratio? e.g. 1 in every 1000 requests Can you replicate the...

Perhaps related to how the forward plugin manages TCP persistent connection pools? @vzemtsov, What version of nodelocalDNS are you using?

Thanks! These kind of docker issues can be mysterious (to me anyways). Can you explain why this fix works? It looks in part similar to a past attempt to fix...