Hello, We are currently managing StarRock [shared-data]which we use it as the query engine part connect to Hive catalog [current load is read data from hive catalog] and I’d like to ask for your advice regarding our Kubernetes cluster design. Current setup • StarRocks on Kubernetes 10 CN servers and 3 FE servers • Per node: 192 CPU cores, 1.5 TB RAM, 14 TB disk • Deployment: 1 pod per node • Pod resources: 172 CPUs and 1.2 TB memory per pod • We set one pod per one node Upcoming changes We are going to receive more nodes [expand cn]: 8 servers with 128 CPU cores and 1 TB RAM , 14 TB disk We are considering the following : • expand the cluster by reduce pod size to CPU 40 core and 384 mem • it will get 4 pods per one old node and 2 pod per one new node • total 56 pods with CPU 40 core and 384 mem Im not sure that it will be the right way to use smaller pods because queries on the data lake are highly memory-intensive. Do you have any idea about my case?