How does SparkGraphComputer work with remote Janus docker?

[melgenek]

Hi, Janus community!

I am setting up a Janus deployment.
My setup is going to have Scylla a backend.
Janus itself is deployed as a set of pods of janusgraph-docker which are exposed via the JanusGraphWsAndHttpChannelizer.

There will be 2 consumers of Janus:

1. an http server that allowes querying nodes by id
2. a spark cluster doing some traversals

I cannot connect to Janus gremlin server with this code

  val cluster = Cluster.build()
    .addContactPoint("my-endpoint")
    .port(8182)
    .serializer(Serializers.GRAPHBINARY_V1D0.simpleInstance())
    .create()
  val remoteConnection = DriverRemoteConnection.using(cluster, "my_traversal")
  val g = AnonymousTraversalSource.traversal().withRemote(remoteConnection)

  g.V().label().next(20)

What's not really clear is where the traversal is actually gonna happen.

My assumption is that the code above would perform the whole traversal on the remote Janus instance that is running in pod, and the complete result is returned to the client.
Am I correct?

If the previous statement is true, it would be nice to tweak the behaviour for Spark. What I'd like to achieve is pulling the data from Scylla on the remote Janus docker but doing the traversal in spark. Will it suffice to add the .withComputer(SparkGraphComputer) to the AnonymousTraversalSource.traversal().withRemote(remoteConnection)?

Thank you in advance for you answers!