When a cluster member does not respond before timeout, other members can fail to join the cluster. Previously, in this case, the logs contained only an entry about the timeout event that occurred while waiting for responses from the cluster ("GMS flush by coordinator at <ip_address>:<port> failed ") and did not report which members were not present in the cluster. Now, the log level of this event was changed from DEBUG to WARN, and the log event contains a list of the missing cluster members to allow easier troubleshooting.