As a part of
SERVER-48347, we added debug logging to the isSelf codepath to determine why isSelf could fail.
However, after doing so, we noticed that other suites would also benefit from this logging. Since they are logged under the NETWORK component, we would have to increase the verbosity per suite (sometimes in reaction to seeing a failure we could not diagnose without extra isSelf logging).
It could be useful to change these logs to warnings or log them at the default log level. Based on the conversation in BACKPORT-7368, we should investigate the risk of flooding the logs when a node is removed and it retries isSelf on every heartbeat. It could end up being too verbose.