How to scale Rasa X to handle more concurrent users

[Deprecated] Rasa X Community Edition
1

Rasa X: 0.39.2 Rasa: 2.6.0-full

Server:

4 Core 16GiB Ram 150GB SSD

During testing we saw one of the rasa x components become CPU bound. Which limits performance. I do not properly recall which component.

The main questions to the rasa x team is: How to scale Rasa X to handle more concurrent users?

We have use Helm to install Rasa X on Kubernetes. Which deployments or statefulset needs to be scaled up? And does this require us to reconfigure anything else besides scaling out more pods?

Thanx in advance!

2

You can increase the number of replicas for your pods. In my opinions, these are the most useful pods to replicate:

Service Role
rasa-x Running the HTTP API
rasa-production Running a trained model, parsing intents, predicting actions
rasa-worker Training and evaluating models

You can learn more in the Rasa Advanced Deployment Workshop .