Breadcrumbs

RASA-X Deployment

Deployment prerequisites

Type

RAM (GB) 

CPU

DISK

Scalability 

Network  Ports

Minimum Nodes

Single-Node

Multi-Node

Single-Node

Multi-Node

Single-Node

Multi-Node

RKE21

8

8

250 GiB 

250GiB (preferably on /var)

No

High

  • 6443/TCP and 9345 to be accessible by all nodes

  • 8472/UDP for CNI

  • 10250/TCP for metrics-server

  • 2379-2380/TCP for Cluster HA (Ports info can be found here)

1

2+


RASA-X 

Rasa-X deployment on same node is not recommended . Install it on a separate node to deploy a production ready system. Deploying on the same node severely degrades the overall performance

  1. Create the rasa-x namespace using

    kubectl create namespace rasa-x

  2. Add the image pull secrets for rasa-x

    kubectl apply -f pre-deployment/registryCredits/ef-imagePullSecret-rasa-x.yaml

  3. Add the Nginx customized config for RASA-X as Confirg-Map

    kubectl create -f pre-deployment/rasa-x-1.1.2/ef-rasa-x-nginx-standard-conf.yaml

Update the FQDN parameter for RASA-X

  1. Change the <CIM-FQDN> with your actual FQDN for CIM Solution.

    sed -i -e 's@value: http://devops.ef.com/bot-framework@value: https://<CIM-FQDN>/bot-framework@' external/rasa-x/values-small.yaml

  2. And

    sed -i -e 's@value: https://devops.ef.com@value: https://<CIM-FQDN>@' external/rasa-x/values-small.yaml


  3. Install the rasa-x using helm command

    helm  upgrade --install=true --wait=true --timeout=10m0s --debug   rasa-x    --namespace rasa-x     --values external/rasa-x/values-small.yaml  external/rasa-x

Rasa-x deployment may take longer than expected due to dependency on huge image sizes and may throw warning like Error: timed out waiting for the condition  , however the deployment continues. You can check the status of the deployment by running ` kubectl get pods -n rasa-x ` which shall give more recent information.

  1. You RASA-X service is available through  HTTP 30800 NodePort on all the nodes' IP Addresses in your Cluster.