Optimize server performance

Hi, I'm running an Elasticsearch cluster and I'm not sure we're maximizing the usage of server resources. When bulk-indexing at maximum speed, CPU, disk I/O and garbage collection seem all pretty low, yet Elasticsearch rejects new bulk requests because the thread pool is full. I'm suspecting the limit might be on some configuration setting. Any ideas on which resource to look up or which setting to tweak to see if performance can be improved? Basically, find out what's the bottleneck? Thanks!

https://www.elastic.co/cn/blog/why-am-i-seeing-bulk-rejections-in-my-elasticsearch-cluster This can help you.

I'd also try looking at the hot threads API to see what it's so busy doing. For instance, if the bottleneck is slow IO then you might see low CPU and low disk IO, but you may see lots of threads waiting on IO operations, particularly things like a force() (i.e. a fsync() call).


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