I have Elastic Stack v7.4 installed on cloud. I have 8 nodes (3-master, 3-Hot, 2-Warm)
ILM policy is in place which keeps one month recent data in hot node and the remaining in warm data node.
Now when I do a query (basically loading discover page), CPU on both warm nodes shoots up to 100% and latency goes upto 25000ms, resulting in time out.
I have 40 shards (1-primary, 1-replica) on my cluster. JVM heap utilization is 2gb/8gb (constant) on warm node. Data is around 100M events.
Can someone please help me understand what may the cause of such behaviour. Also why CPU is used totally and RAM is low and constant.