Index rate peaking at 4 p\s on index of 2.2m documents

I have an index of 2.2m records. My index rate is at best 4 per second and periodically 0 ps.

The settings are below. Any insight gratefully received.
"dev-es-cdb": {
"settings": {
"index": {
"refresh_interval": "1s",
"number_of_shards": "5",
"provided_name": "dev-es-cdb",
"merge": {
"scheduler": {
"max_thread_count": "10",
"max_merge_count": "10"
"policy": {
"max_merge_at_once": "10",
"max_merged_segment": "10gb",
"segments_per_tier": "25",
"floor_segment": "100mb"
"creation_date": "1558616782785",
"number_of_replicas": "0",
"uuid": "s6ZzNoubQge_vcZvQMXWFg",
"version": {
"created": "5050099"


please take some more time to properly describe your setup, like Elasticsearch version, JVM version, operating system, the speed of your disks with a benchmark tool, what kind of disks/filesystem is on the data path, number of nodes, heap configuration. See

Also, I would personally refrain from trying to tune merge scheduler and policies and try with an empty setup first. Can you explain how you got to those numbers that are listed in your configuration?


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