ES query operation’s impact on cluster CPU consumption

Elastic Search
1

Hello, I have a practical question:

Background: Query operations on ES in stress testing scenarios

Question: The CPU usage of a specific node is much higher than that of other nodes. How to evenly distribute query operations to the ES cluster CPU usage in this scenario?

index
    number_of_shards: 5
    number_of_replicas: 1
es cluster node number 5

image
image428×143 16 KB

2

Which version of Elasticsearch are you using?

Is the data distributed evenly across the nodes?

How many concurrent queries are you running?

Are concurrent queries distributed evenly across the cluster?

What does the hot threads API show for the node under heavier load? Anything that differs compared to the output from the other nodes?