It means that Elasticsearch is not able to keep up with the number of
requests you are sending and that the internal queues are filling up. A
discussion around this can be found in the following blog post:
Best regards,
Christian
On Thursday, April 30, 2015 at 8:58:36 AM UTC+1, xinm...@163.com wrote:
When doing pressure test on ES cluster,it sometimes may return 429(too
many request).
I can't find any detail docs about this.In which case Elasticsearch may
return 429 ?
However, I still feel confuse because when we keep 200 concurrent thread
cyclically send requests (Indexing request) to cluster,it performence good
and we guess there still have remaining performance space.You can see the
"Throughput" is 5532.8/sec.
As explained in the blog post, increasing the queue size will not improve
performance, just make you store more data in memory on the cluster
awaiting processing. It could actually instead end up reducing performance.
It looks like you are hitting the limit of your cluster and that the number
of concurrent indexers it is able to keep up with is somewhere between 200
and 300.
Have you been able to identify what is limiting performance when the queues
start building up and Elasticsearch can't keep up?
Christian
On Thursday, April 30, 2015 at 8:58:36 AM UTC+1, xinm...@163.com wrote:
When doing pressure test on ES cluster,it sometimes may return 429(too
many request).
I can't find any detail docs about this.In which case Elasticsearch may
return 429 ?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.