# ES performance improvement after restarting ES instance?

**URL:** <https://discuss.elastic.co/t/es-performance-improvement-after-restarting-es-instance/340564>\
**Category:** Elastic Observability\
**Created:** [August 10, 2023, 1:29pm UTC](https://discuss.elastic.co/t/es-performance-improvement-after-restarting-es-instance/340564 "2023-08-10T13:29:12Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![hyt](https://avatars.discourse-cdn.com/v4/letter/h/f08c70/32.png) [@hyt](https://discuss.elastic.co/u/hyt)\
**Post date:** [August 10, 2023, 1:29pm UTC](https://discuss.elastic.co/t/es-performance-improvement-after-restarting-es-instance/340564/1 "2023-08-10T13:29:12Z")

</div>

I encountered a strange phenomenon where the Elasticsearch instance performed better after restarting it during performance testing with esrally. i don't know why?  
esrally ：v2.6.0 （official http\_logs track）  
es : v7.17.9 （only one es instance）  
x86\_64 ,16 cpu, 64G mem( jvm use 31G)  
step:  
1.

```auto
esrally race --track="http_logs" --target-hosts="$server_hosts" --pipeline=benchmark-only --kill-running-processes

```

1. because I am only interested in range queries. When the target-throughput of range queries is set to 230 with 16 clients, the CPU load is at 100% capacity.

```auto
esrally race --track=http_logs --target-hosts="$server_hosts" --pipeline=benchmark-only --kill-running-processes --offline --include-tasks="range"

```

1. restart es instance  
4.Running the following command with the same target-throughput and clients，the CPU load is only at 50% capacity, but the performance results output by esrally are consistent with the results obtained at full load.

```auto
esrally race --track=http_logs --target-hosts="$server_hosts" --pipeline=benchmark-only --kill-running-processes --offline --include-tasks="range"

```