Gremlin driver best practices for closing cluster and client objects

Question

I am using gremlin-driver to connect to cosmosDB. I am working with multiple graph instances simultaneously. I am storing the client objects in a cache, and using it to submit gremlin queries. How should I go about closing these on cache evict, when the maxTtl limit is hit? I am planning to call client.getCluster().close() on cache evict, but not sure what happens to the existing queries in flight when the close is called. What is the best practice here?

Answer 1

This is how we are solving the problem. We are adding the evicted clients to a secondary cache with that evicts after a predefined interval(30s). This is to give sufficient time for the pending transactions to complete. On the secondary evict, we close the client and cluster.