Is there an API to get the read and write performance data?

[KeithTt]

ES version: 6.3.0

My ES cluster block sometimes when perform search and bulk operations, but the cluster status is green, it is strange.

The cat.thread_pool api can output some related info, but i do not know how to determine the abnormal status.

node_name name                active queue rejected
m-21-63   analyze                  0     0        0
m-21-63   fetch_shard_started      0     0        0
m-21-63   fetch_shard_store        0     0        0
m-21-63   flush                    0     0        0
m-21-63   force_merge              0     0        0
m-21-63   generic                  1     0        0
m-21-63   get                      0     0        0
m-21-63   index                    0     0        0
m-21-63   listener                 0     0        0
m-21-63   management               1     0        0
m-21-63   ml_autodetect            0     0        0
m-21-63   ml_datafeed              0     0        0
m-21-63   ml_utility               0     0        0
m-21-63   refresh                  0     0        0
m-21-63   rollup_indexing          0     0        0
m-21-63   search                   1     0  1946236
m-21-63   security-token-key       0     0        0
m-21-63   snapshot                 0     0        0
m-21-63   warmer                   0     0        0
m-21-63   watcher                  0     0        0
m-21-63   write                    0     0    12383

Is there a proper way to get or compute the performance information?

[warkolm (Mark Walkom)]

Do you have Monitoring enabled?

[KeithTt]

Yes, i can see the monitor page.

[warkolm (Mark Walkom)]

What does it show when this behaviour happens? High GC, threadpool use, memory/CPU use?

[KeithTt]

I just can see the cup/memory/disk and search/index rate.

Where is the GC and thread pool info?

[KeithTt]

Maybe I can not see these info because the basic license?

[warkolm (Mark Walkom)]

Check out View monitoring data in Kibana | Kibana Guide [7.11] | Elastic

[KeithTt]

My ES and Kibana version is 6.3.0, and I do not use metric beat.

I have two cluster, and I have compare the two versions, 6.3 and 7.11, there is no GC and thread pool info both of them.

Did I miss anything?

[warkolm (Mark Walkom)]

6.3 has been EOL for quite some time, you should upgrade ASAP.

In 7.11 you need to go to Stack Monitoring > choose your cluster > Nodes > pick your node, then you will see GC.

[KeithTt]

Thanks a lot. I find the GC graph.

And how do I get the thread pool info and something else about performance...by an API?