I’m interested in setting up a Rasa Core server with a single model but multiple independent conversations. I had a dig around and it seems the best option is to have a different tracker store (say mongodb) for each new conversation. Is it correct that this, along with different endpoints, is all that is required in order to compartmentalise the conversations entirely? This would be great as it’s too big to fit in an AWS Lambda and spinning up lots of docker containers would be a bit slow (and it’s expensive to maintain unused spares all the time).
Also open to other options/suggestions, thanks!