Elasticsearch thread_pool is growing over and over

Hi all,

I am a newbie with Elastic. I cannot find all my documents in my research. If I repeat the same search, after a while, it's okay. I understood that it could come from the 'refresh'. But sometimes the waiting time is really long. I have found that the 'thread_pool' is getting bigger and bigger. I do not know if it's normal..

Any ideas? Please

Version 7.8.0 on Windows Server 2019.

Refresh is happening behind the scene "when it's needed".

Could you explain a bit more about the context? Is that for integration testing? Is that in production?
What is the problem of "waiting a bit" for your use case?

It's in a production context.

I will try to explain my use case.
I index documents on the fly, 10K per day, mostly on working hours. Indexing, one by one not in batch.
Also, I search the index before indexing the document. So the refresh is normally triggered regularly.
Maybe an hour after indexing one, I'm looking for it (this one in particular) for the first time and sometimes I can't find it.
All this with the Nest client (in netcore).
If after that, I redo the same request directly with REST and the document is found.

This is why I am talking about refresh and why I am wondering about the queue.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.

It should not really happen, after one hour.
In case it helps, you can trigger manually a refresh just before running your search. See Refresh API | Elasticsearch Guide [7.12] | Elastic