I’m able to deploy RASA-X on local and as well as on an AWS instance using both docker-compose and kubernetes based deployments. I’m sharing the screenshots for the details on running instances for kubernetes and docker-compose.
There is no documentation on how these containers / pods interact with each other. For instance, I’d like to understand in detail on how the pods/container interact with each other when I use the following RASA-X REST API. I’m using a custom action server.
@KarthiAru glad to see you’ve managed to get it deployed. That’s definitely a very valid question you have, right now we don’t have a publicly available diagram, however we’re planning on adding that soon.
api/chat will hit the rasa-x container, which will forward the request to the
rasa-production container, which takes care of making a rasa prediction (nlu/core). That will then hit the
app container if there’s a custom action to be executed, if there isn’t it’ll send the response back to
rasa-x to send back to the user
We are going to have an technical Rasa presentation in our company. I would be very glad, if you could publish an architecture visualization and description of Rasa X, as it gets installed as described here. We need it in 3 to 4 weeks. Thank you so much.