Before Oct 18, static 99% less than 439ms, After Oct 18 is 603ms.
cluster: 5 node mdi, 32core 128G memory(but jvm config is 31g). 1 index 100shards,2b documents with 2T disk used space.
Is it because there's too much documentation?
Disk I/O is often a limiting factor in Elasticsearch. What does it look like in your cluster, e.g. using iostat? Has your data set grown or changed?
Thanks.
I find that the I/O usage rate is indeed soaring.
Could you tell me how to improve this situation?
Add more nodes?
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.