Separate write and read operations between different Trino workers

[malcolm061990]

Hi team

Is there any mechanism in Trino (kind of "roles") to separate write and read queries between specific workers using single coordinator?

Our use case:

• write pipeline is critical. So we use on-demand k8s node group for these queries
• read pipeline is highly scalable due to queries from different apps. So we use Spot k8s node group for these queries

Or is it better to deploy two separated Trino clusters (one for read, another one for write) using same Nessie and send queries depending on their type to required cluster?