Tranquility internals : zookeeper connection


I have a query about tranquility-zookeeper communication.

when using tranquility core API zookeeper URL is configured in the tranquility configuration (zookeeper.connect). Then whenever data is to be sent to druid through tranquility, service discovery happens with tranquility figuring out the overlord addresses, firehose, segments announced etc. from zookeeper. This all works seamlessly allowing data ingestion into druid.

But what happens when we have 2 overlord and 2 coordinator nodes which works in failover configuration? So, first tranquility would be able to find the overlord/coordinator address and send data but does tranquility caches these addresses or would it be able to handle the failover of a overlord/coordinator? i.e. when the coordinator/overlord leadership changes.