During upgrade/downgrade testing, errors are encountered while the apiserver is down. This is normal and handled via retrying, so we don't need to be verbose.