Detect Failover of MongoDB-Cluster with Spring-Data-MongoDB

Question

Current Situation

• we have a MongoDB-Cluster with 1 primary node and 2 secondary nodes
• our Spring-Boot application is using the Spring-Data-MongoDB framework to read/write from/to the cluster

Problem

• in some circumstances the MongoDB cluster will change the primary node (for example during the resizing of the cluster)
• this fail-over phase will affect our Spring-Boot application
• when some reads or writes are still ongoing and the fail-over happens, we receive an exception, because the mongoDB-Server is not reachable anymore for our application
• we have to deal with this state somehow

Questions

1. What is the best way to handle those faile-over states ?

• I've come across the following documentation:
  • retryable writes
  • retryable reads
• would it be sufficient to set the retryReads and retryWrites flag to true and specify the primary node and the secondary nodes in the connection url? Or should we catch the connection-exception (or alternatively listen to some fail-over-event) and handle those cases by ourself ?
• we also have to deal with the following problem: what happens if only 50 % of some bulk-write data got successfully written to the primary node and the other 50 % not ? How handle those cases ideally ?
• this leads us to the second question ...

2. How to detect the fail-over event in Spring-Boot ?

• for our application a possible solution would be to automatically detect the failover state of the MongoDB-Cluster and than just trigger a restart of our Spring-Boot application.
• is there a way to listen to a specific MongoDB-event via spring-data-mongodb in order deal with the case that the primary node has changed?
• alternatively: is there a specific exception we should catch and handle?

I hope somebody can help us here. Thank you in advance!