Disable DNS caching in the Java Virtual Machine - Amazon Neptune
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Disable DNS caching in the Java Virtual Machine

In an environment where you want to load-balance requests across multiple read replicas, you need to disable DNS caching in the Java Virtual Machine (JVM) and provide Neptune's reader endpoint while creating the cluster. Disabling the JVM DNS cache ensures that DNS is resolved again for every new connection so that the requests are distributed across all of the read replicas. You can do this in your application's initialization code with the following line:

java.security.Security.setProperty("networkaddress.cache.ttl", "0");

However, a more complete and robust solution for load-balancing is provided by the Amazon Gremlin Java client code on GitHub. The Amazon Java Gremlin client is aware of your cluster topology and fairly distributes connections and requests across a set of instances in your Neptune cluster. See this blog post for a sample Java Lambda function that uses that client.