I've found that proxysql also has issues with the Aurora bug.
When clone or instance name is renamed in Aurora, the instances of the source cluster and the instance name before the change are included in the 'information_schema.replica_host_status' table result.
This is recognized as a bug by AWS, but there is no fix schedule yet.
When searching the 'information_schema.replica_host_status' table in ProxySQL, the condition LAST_UPDATE_TIMESTAMP >= DATE_ADD(NOW(), INTERVAL -1 MINUTE) must be included.