[
https://issues.apache.org/jira/browse/UNOMI-237?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16882151#comment-16882151
]
Serge Huber commented on UNOMI-237:
-----------------------------------
I was thinking of maybe increasing the size of the pools for indexing :
https://www.elastic.co/guide/en/elasticsearch/reference/5.6/modules-threadpool.html
And maybe freeing up some CPU for ElasticSearch by moving it to another machine
would be a good idea if that's possible (could be limited to the import
operation).
Basically it would be good to check if your are CPU or IO bound and see if you
can address those problems during the import.
Regards,
Serge...
> 204 status code when under high traffic
> ---------------------------------------
>
> Key: UNOMI-237
> URL: https://issues.apache.org/jira/browse/UNOMI-237
> Project: Apache Unomi
> Issue Type: Bug
> Components: core
> Affects Versions: 1.4.0
> Reporter: Michele Riva
> Priority: Major
>
> Hi all,
> I've noticed that after spawning 10.000/20.000 consecutive POST requests to
> /cxs/profiles/create, Unomi begins to return a 204 status code and the
> profile can not be created.
> I've acted adding a retry strategy on my client, but obviously it slows down
> the system.
> Seems that Unomi can not handle an huge amount of consecutive/concurrent
> requests without an event broker (Kafka/Redis/Rabbit) before it... is there a
> way to guarantee that it will work under high traffic without a broker
> handling the request queue for it?
> thanks!
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
