I’m using a shared_data mode cluster on EKS and using an hpa on the CN. while i see the autoscaling working well, i noticed that when a new pod is created, the query that’s running on the existing CNs (before the autoscaling started) is not distributed to the new CN is it something that can be changed? would it be possible that the added resources would be utilized?
the existing running query can’t be tuned while new CN nodes are added. new queries can be benefited from the new CN resources. However it will be related to your table design, how many buckets are chosen for the partition.