Dear everyone,
Is there any guideline on how to configure the appropriate concurrent number of queries and reflection creations in the cluster?
Any help is appreciated.
Hi @borasy
Depends on the number of executor, capacity of these executors and most importantly type of workload, say you run one high cost reflection and you have 4 executors, if this reflection job uses 30 GB of direct memory (open profile and expand nodes under the query tab) per node and you have about 75 GB of direct memory on that executor then you can set concurrency as 2