Feature Request

What problem are you trying to solve?

Currently, when a DNS query to canonicalize a name fails, the proxy logs a message at the error level. If there is a previously successful canonicalization, the proxy will fall back to that.

In some cases, such as timeouts, these errors are pretty transient, but may occur quite frequently. This results in a lot of error messages being logged but may not cause any actual failures (see also #2069 (comment)).

How should the problem be solved?

When the proxy has fallen back successfully, it might be better to log a warning, rather than an error.
We might also want to consider downgrading timeouts to warnings. We could also change the canonicalization code to log only on state changes, rather than logging every individual timeout.