we have a production cluster based on 4 data nodes used by our frontend application.
We have recently developed a feature that consumes a lot of resource from this cluster, and when one is using this feature, the other users are impacted with low performance (due to lack of space within the search pool).
So I wonder how to handle without to have two separate clusters (one for this feature and another for the rest of the application) ?
Is there a way to create a dedicated node with all the shards replicated, and to route the requests to this node ?
By the way we are using a quite old ES version (2.0)
Thank you in advance,