The sdam spec requires that ismaster calls are retried (https://github.com/mongodb/specifications/blob/master/source/server-discovery-and-monitoring/server-discovery-and-monitoring.rst#retry-ismaster-calls-once). We generally log retries elsewhere, but not in the monitoring thread as evidenced by 1713. We should also log ismaster retries on monitoring threads as this will assist in troubleshooting server selection failures.