# Opensearch performance and throughput boundaries

**URL:** <https://forum.opensearch.org/t/opensearch-performance-and-throughput-boundaries/7832>\
**Category:** Performance Analyzer\
**Created:** [December 1, 2021, 10:02am UTC](https://forum.opensearch.org/t/opensearch-performance-and-throughput-boundaries/7832 "2021-12-01T10:02:58Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![mann](https://avatars.discourse-cdn.com/v4/letter/m/91b2a8/32.png) [@mann](https://forum.opensearch.org/u/mann)\
**Post date:** [December 1, 2021, 10:02am UTC](https://forum.opensearch.org/t/opensearch-performance-and-throughput-boundaries/7832/1 "2021-12-01T10:02:58Z")

</div>

I am trying to get an answer to how much event ingestions/ queries can be done on OpenSearch per unit time? As I am trying to use OpenSearch which will be shared among multiple users then it will be able to handle multiple requests/queries/operations at a single time? If yes what will be the limits, such that all things will work smoothly?

---

<div class="post-metadata">

**Author:** ![reta](https://avatars.discourse-cdn.com/v4/letter/r/cdc98d/32.png) [@reta](https://forum.opensearch.org/u/reta)\
**Post date:** [December 1, 2021, 2:50pm UTC](https://forum.opensearch.org/t/opensearch-performance-and-throughput-boundaries/7832/2 "2021-12-01T14:50:10Z")

</div>

I would recommend you to take a look at Rally ([https://github.com/elastic/rally](https://github.com/elastic/rally)), it works with Opensearch clusters as well most of the time.

---

<div class="post-metadata">

**Author:** ![robcowart](https://sea1.discourse-cdn.com/flex019/user_avatar/forum.opensearch.org/robcowart/32/1012_2.png) [@robcowart](https://forum.opensearch.org/u/robcowart)\
**Post date:** [December 7, 2021, 8:17am UTC](https://forum.opensearch.org/t/opensearch-performance-and-throughput-boundaries/7832/3 "2021-12-07T08:17:17Z")

</div>

This will be very specific to the use-case. For example high-ingest rate logging has a completely different resource requirements than high-query load website search. You will need to share a bit more information about how the cluster will be used.
