How to Integrate IBM Watson with Kubernetes

Prerequisites

• Ensure you have a Kubernetes cluster running. You might use local solutions like Minikube or cloud providers like GKE, EKS, or AKS.

 

• Install the IBM Cloud CLI and the IBM Cloud Kubernetes Service plug-in.

 

• Set up an IBM Watson service instance with the appropriate credentials.

 

• Ensure `kubectl` is installed and configured to interact with your Kubernetes cluster.

Setup IBM Cloud CLI

• Log in to IBM Cloud using the CLI:

ibmcloud login

• Install the Kubernetes Service plug-in:

ibmcloud plugin install kubernetes-service

• Access your Kubernetes cluster:

ibmcloud ks cluster config --cluster <cluster-name-or-id>

Create a Watson Service on IBM Cloud

• Navigate to IBM Cloud catalog and select the Watson service you need (e.g., Watson Assistant).

 

• Follow the instructions to create the service instance and note the credentials provided.

Creating a Kubernetes Secret for Watson Credentials

• Create a secret in Kubernetes to securely store your Watson service credentials:

kubectl create secret generic watson-credentials --from-literal=username=<your-username> --from-literal=password=<your-password>

Deploying Watson API on Kubernetes

• Create a deployment YAML file to define the Watson API setup. Include environment variables to consume the Kubernetes secret:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: watson-deployment
spec:
  replicas: 1
  selector:
    matchLabels:
      app: watson
  template:
    metadata:
      labels:
        app: watson
    spec:
      containers:
      - name: watson-container
        image: <your-watson-image>
        env:
        - name: WATSON_USERNAME
          valueFrom:
            secretKeyRef:
              name: watson-credentials
              key: username
        - name: WATSON_PASSWORD
          valueFrom:
            secretKeyRef:
              name: watson-credentials
              key: password

• Apply the YAML to deploy it:

kubectl apply -f watson-deployment.yaml

Expose the Watson API

• Create a service definition file to expose the Watson API deployment:

apiVersion: v1
kind: Service
metadata:
  name: watson-service
spec:
  selector:
    app: watson
  ports:
    - protocol: TCP
      port: 80
      targetPort: 8080
  type: LoadBalancer

• Apply the service configuration:

kubectl apply -f watson-service.yaml

Verify the Integration

• Get the external IP of the service by running:

kubectl get svc watson-service

• Use this external IP to send requests to your Watson API and verify if it's working as expected.

 

• Ensure your application is configured to route requests to this service.