I’m looking to both run Rasa core/nlu for a facebook bot AS WELL as be able to be able to probe the NLU to get the list of intents for a phrase (used to be /parse on the RASA NLU endpoint default port 5000).
I was able to meet this use case only by launching a container for NLU functionality and a container for Core functionality (same container twice now that they’re joined)
If someone has a better solution without separate containers would be happy to hear it.
In theory perhaps something like supervisord could run two rasa processes in the same docker container with different arguments. Not sure if there are negative repercussions to that.