When zookeeper recovers from the interruption， How can I ensure that vertxcluster remains robust？

When zookeeper recovers from the interruption, vertx throws an exception. Do I need to switch again? Is the previously registered bus still valid? How can I ensure that vertxcluster remains available in the event of zookeeper interruption and recovery?  
i set curtor retry forevery to make sure connect  zookeeper when its alive，but
vertxcluster seems  cant recover from a zookeeper  collapse