Rasa X deployment issue with custom action server

[Deprecated] Rasa X Community Edition
1

I have been following the deployment methods mentioned here to deploy a bot into production, Ubuntu 18.04 using CLI.

So far I have completed the quick install script and pushed the custom server image into dockerhub. The part where I am stuck is Adding a Custom Action Server to the values.yml file.

  1. I am not sure where to find the file itself so I cannot update the values there manually
  2. helm repo update helm upgrade
    –values values.yml
    –namespace rasa
    –reuse-values
    rasa rasa-x/rasa-x

This command fails with:

Error: UPGRADE FAILED: template: rasa-x/templates/rasa-x-deployment.yaml:57:45: executing “rasa-x/templates/rasa-x-deployment.yaml” at <.Values.rasax.livenessProbe.scheme>: nil pointer evaluating interface {}.scheme

1 Like
2

Hi, It looks like your deployment uses an older helm chart version than you trying to use. Additionally, your value.yml is probably out-of-date.

You have several options:

  • Be sure that you use the same version of the helm chart during the upgrade that your deployment use
  • Don’t use --reuse-values
  • Be sure that your values.yml files include missing fields from the latest version
1 Like
3

Hello @tczekajlo Thanks for taking the time to reply. The deployment piece is a little complicated and requires tech stack I am not familiar with.

I have installed rasa-x from here using quick install script: curl -s get-rasa-x.rasa.com | sudo bash

export KUBECONFIG=/etc/rancher/k3s/k3s.yaml

Now the quick install script does 3 things as per the doc:

  1. Install an embedded Kubernetes cluster.
  2. Install the Helm command-line interface.
  3. Install Rasa X using the Rasa X Helm Chart.

So all helm chart related configuration was done by the script itself and I kept it that way as I was not familiar with it.

Rasa-x is running and I am able to access it via browser. The custom action server is pushed to docker and that is something which I want the bot connect.

image
image791×43 16.3 KB

helm list shows nothing running

kubectl get pods --all-namespaces

image
image936×361 171 KB

What do I do next?

helm version

version.BuildInfo{Version:“v3.4.0”, GitCommit:“7090a89efc8a18f3d8178bf47d2462450349a004”, GitTreeState:“clean”, GoVersion:“go1.14.10”} `

4

I got a similar error and removing the --reuse-values option solved the problem. Thanks