High load, but lower cpu,lower IO

[111421 (QQ)]

Hello, eyeryone
I checked the es single node load through cerebor and found that the load is very high, but the cpu utilization rate is very low. Why?

ES is deployed through docker, version: 6.8

cerebro:

kibana:

GET /_nodes/hot_threads

::: {es-node1}{VxAcSdn8QJKRBu705pFVDA}{lOWzUj-8SVKTo3X_Tku42Q}{10.244.0.15}{10.244.0.15:9300}{ml.machine_memory=66933485568, xpack.installed=true, ml.max_open_jobs=20, ml.enabled=true}
   Hot threads at 2020-12-18T02:31:16.835Z, interval=500ms, busiestThreads=3, ignoreIdleThreads=true:

    0.0% (140.2micros out of 500ms) cpu usage by thread 'elasticsearch[es-node1][refresh][T#1]'
     4/10 snapshots sharing following 2 elements
       java.base@13.0.1/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
       java.base@13.0.1/java.lang.Thread.run(Thread.java:830)

[rugenl (Len Rugen)]

Check your iowait

Linux Load Averages: Solving the Mystery

[111421 (QQ)]

iostat -x 1:

[rugenl (Len Rugen)]

Sorry, I missed that you're under Docker. You probably need to look at docker stats instead.

[111421 (QQ)]

docker stats:

Load is still very high now

I am not sure if it is related to deployment in docker

[rugenl (Len Rugen)]

I think docker stats are "outside looking in", so low CPU usage probably reflects low demand for cpu.

I don't recognize where you are getting the load stat, but if it is from INSIDE the container, I'd question it's validity. If you don't have performance problems, don't worry about it.