We are building a bot which will encounter 100s of messages simultaneously
Q1) Is it recommended to use API Gateway while deploying Rasa? Why and Why not?
Q2) Should we consider switching to k8s from k3s for deployment? (Given the bot will face high traffic)
We recommend our Kubernetes Helm configuration for high traffic/availability environments. This can be deployed on your favorite Kubernetes service.
I recommend this over your own API gateway configuration because our engineers maintain and update this with new features and new releases of Rasa.
You can also customize this to your requirements if you have deployment preferences or IT standards.