# Rasa in production (Performance Benchmarking and how to scale)

**URL:** https://forum.rasa.com/t/rasa-in-production-performance-benchmarking-and-how-to-scale/21559
**Category:** Rasa Open Source
**Created:** [November 21, 2019, 5:47pm UTC](https://forum.rasa.com/t/rasa-in-production-performance-benchmarking-and-how-to-scale/21559 "2019-11-21T17:47:48Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![shubham3582](https://dub1.discourse-cdn.com/flex013/user_avatar/forum.rasa.com/shubham3582/32/5872_2.png) [@shubham3582](https://forum.rasa.com/u/shubham3582)
#### Post date: [November 21, 2019, 5:47pm UTC](https://forum.rasa.com/t/rasa-in-production-performance-benchmarking-and-how-to-scale/21559/1 "2019-11-21T17:47:48Z")

</div>

Following questions still are in my mind before we are shipping our rasa based solution to production:

1. Is there any performance benchmarking for rasa core and rasa nlu
2. how we can scale rasa (both horizontally and vertically)
3. Is it possible to share any production grade architecture of rasa based solution?
4. Since we might be getting very high workload (50000 concurrent user), any suggestions apart from upgrading it to rasa-x?
