How to deploy duckling next to rasa x?

I want to use rasa x to make my chatbot available to an external audience. I used Microsoft Azure to create a virtual machine and use the server quick install manual. Worked like a charm: rasa x is approachable from the outside and starts every time the server reboots.

My chatbot uses duckling to extract numbers. Following the instructions in the doc I am on my local machine able to run the docker with duckling and the bot is able to communicate with it.

How do I do this in the cloud? Do I run the docker image parallel to rasa x or create a new virtual machine? How do I make sure that the docker image runs whenever the system boots? What is the correct url to connect to duckling - is it still considered localhost?


The server quick install script has an environmental variable, ENABLE_DUCKLING that will automatically start a Duckling pod as part of your deployment. See the customization docs here: Customize Your Deployment