?? Deep Dive into Implementing Shadow Testing with Diffy in Kubernetes

Following up on the importance of Shadow Testing in Kubernetes using Diffy, let's explore the practical side of this innovative testing strategy with some code examples!

???? Step-by-Step Guide to Implementation

1.Setting Up the Environments:

V-Current Deployment

apiVersion: apps/v1
kind: Deployment
metadata:
  name: v-current-deployment
  namespace: production
spec:
  replicas: 2
  selector:
    matchLabels:
      app: myapp
      version: v-current
  template:
    metadata:
      labels:
        app: myapp
        version: v-current
    spec:
      containers:
      - name: myapp
        image: myapp:v-current
        ports:
        - containerPort: 80        

V-Next Deployment

apiVersion: apps/v1
kind: Deployment
metadata:
  name: v-next-deployment
  namespace: testing
spec:
  replicas: 2
  selector:
    matchLabels:
      app: myapp
      version: v-next
  template:
    metadata:
      labels:
        app: myapp
        version: v-next
    spec:
      containers:
      - name: myapp
        image: myapp:v-next
        ports:
        - containerPort: 80        

2. Deploying Diffy:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: diffy
spec:
  replicas: 1
  selector:
    matchLabels:
      app: diffy
  template:
    metadata:
      labels:
        app: diffy
    spec:
      containers:
      - name: diffy
        image: opendiffy/diffy:latest
        ports:
        - containerPort: 8880        

3. Services for Exposure:

• V-Current Service

apiVersion: v1
kind: Service
metadata:
  name: v-current-service
  namespace: production
spec:
  selector:
    app: myapp
    version: v-current
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80        

V-Next Service

apiVersion: v1
kind: Service
metadata:
  name: v-next-service
  namespace: testing
spec:
  selector:
    app: myapp
    version: v-next
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80        

Diffy Service

apiVersion: v1
kind: Service
metadata:
  name: diffy-service
spec:
  selector:
    app: diffy
  ports:
  - protocol: TCP
    port: 8880
    targetPort: 8880        

4. Directing Traffic:

Using the Kubernetes Gateway API for traffic splitting, you can set up an HTTPRoute to distribute a percentage of the traffic to different services. Below is an example where 20% of the traffic is routed to the Diffy service, and the remaining 80% is routed to the v-current-service.

apiVersion: networking.x-k8s.io/v1alpha1
kind: HTTPRoute
metadata:
  name: my-http-route
spec:
  hostnames:
    - "myapp.example.com"
  rules:
    - matches:
        - path:
            type: Prefix
            value: /
      forwardTo:
        - serviceName: diffy-service
          port: 80
          weight: 20
        - serviceName: v-current-service
          port: 80
          weight: 80        

?? Automation for Efficiency

Consider automating this entire process as part of your CI/CD pipeline. This approach not only saves time but also ensures consistency and reliability in your deployments.