WRITE threadpool rejected stats

Hello all, I'm using Elasticsearch 7.6.1 - I was looking at the indexing stats with the following command:

GET /_cat/thread_pool/write?v&h=id,name,active,queue,rejected,completed

Which gives the following output:

id                                                       name  active queue rejected completed
ch-dgJaQS7mT5cIfsxB7PA                                   write      0     0   7838     355028363
Hz4uMmZdTQSY11W--ZpaDw                                   write      0     0   7695     354988134
nHjK8IN-Rxi2B-VoH4pcHQ                                   write      0     0   1735     354961006

Is this an unusually high number of rejections?

I'm using Logstash to send the data to Elasticsearch and am occasionally seeing response code 429 (rejected BulkshardRequests) in the Logstash log, however these messages are only occasional and they don't seem to last more than a second so I've assumed everything has just sorted itself out - do I need to look into this more or are these only transient problems that can be ignored?


They sound transient to me, and a few thousand rejections in a few hundred million requests sounds ok. Logstash should react correctly to rejections by backing off and retrying after a bit.

Thanks David, I thought that'd be the case.

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