I use esrally. when i setting throughout to 100 or 200, the Median Throughput are about 100 or 200. but i setting throughout parm more, the mediam throughput still about 240. when i watch the monitor, the load less than 2 and cpu usage less than 40%, but the percentile latency is more bigger than before. my guess is that thread_pool.search.size is causing the bottleneck, is right?
we use aliyun elasticsearch, can not modify thread_pool.search.size, so I can't test my theory
It sounds like you may be hitting the throughput limit at 240. Is this for bulk indexing requests or queries?
Disk I/O is often the limiting factor for Elasticsearch performance. Do you have any monitoring statistics for disk I/O, IOPS and/or disk latency?
Which version of Elasticsearch are you using? What is the size and specification of your cluster? What is the size of your data set?
thank you for you reply. only query test pressure measurement.
es version: 7.10.2
index size: 52G 10shards
cluster setting: 4core 8G memory 3 nodes
momery and disk use about 50-60%, disk latency has no monitor
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.