Kubernetes - Setup


I have been having troubles deploying my chatbot using kubernetes. I have already successfully implemented my bot using docker-compose and would like to scale it for more users, but have not been successful thus far.

I’ve tried to use the following tutorial

and I have also watched the video tutorial with no avail (Ep #12 - Rasa Masterclass) Deploying Rasa X in a Cluster Environment | Rasa 1.8.0 - YouTube

Please let me know if there is another tutorial online that I would be able to follow. I would also appreciate it if there were more tips online for debugging and troubleshooting when it comes to implementing a conversational agent with Rasa on Kubernetes.

Best regards, Robyn

which components of rasa do you want to deploy on K8s?

  • Rasa X
  • Rasa Open Source
  • Event Broker(Rabbit, Kafka)
  • Tracker Store(Redis, any database)
  • Lock Store(Redis)

If it is all of the above, then i think the helm chart of rasa is quite complete and will deploy all the components. I would say you should look into Helm a bit to understand what all the files do mean to get accustomed. for me that was the hardest part is to get accustomed to a way of deploying into K8s. I have previously used kompose to break my docker-compose to k8s deployment files but that way of doing got out-dated and now helm charts are definitely the best way to go about it.

1 Like