How to set Amazon Neptune write/read endpoint in configuration

[codingalone]

It's rather a question:

I'm using Grumlin with Amazon Neptune and facing a problem in the title.

In my current code, Grumlin.configure is called only once in a startup script, so I cannot use write/read instances accordingly. Now I only use the write instance for read queries too.

Is there any good way to configure both write and read endpoint in a single server instance?

[zhulik]

Hey, sorry for such a huge delay, Grumlin currently does not support separate read/write instances as I don't have a cluster with read instances and haven't investigated a possibility of adding this feature.

Do you have any ideas on implementation? We could have separate configuration options for this, something like

Grumlin.configure do |config|
  config.url = "ws://localhost:8182/gremlin" # keep it url for backwards compatibility
  config.read_url = "ws://localhost:8182/gremlin"
end

A new middleware could check what steps are used and if there are no mutating steps it could route the query to the read instance, what do you think? Could you help me with testing it?