Node configuration : 61.0 GiB 8 vCPUs and 8 nodes
3 master nodes and 5 data nodes
thread_pool.search.queue_size: 1000
we have almost 300+ indices and 1 Percolator index. Per day indexing is 400K and this same 400K will be searching in percolator as well. So the CPU usage of 3 data nodes are always 92%
how can we scale the setup for this much data