Currently, the client loops continuously whilst a query is running, sending GET fetch requests. For clients issuing a high volume of long running queries (ETL), these requests add unnecessary load to the Trino cluster and any intermediary proxying systems (lyft-gateway for example).
Can we add a configurable sleep (default to 0 to preserve current behaviour) to the fetch logic?
Currently, the client loops continuously whilst a query is running, sending
GET
fetch requests. For clients issuing a high volume of long running queries (ETL), these requests add unnecessary load to the Trino cluster and any intermediary proxying systems (lyft-gateway
for example).Can we add a configurable sleep (default to 0 to preserve current behaviour) to the fetch logic?