I am building my own docker image to run docker and I am seeing a dramatic decrease in performance. I am using rasa_nlu version 0.12.3 with python 3.6. Running locally on my machine, I am able to achieve 70 requests per second being parsed by rasa nlu. However, when using docker, I am getting about 10 requests per second. Has anybody else seen this issue or know how to fix it?
Hi @spenc53, a container usually does not have access to all the resources of the hosting machine, that is why you see a drecreas in performance. You can try to increase the resources allocated to docker in you docker preferences and see if this helps with the performances.