deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-flask-app
  labels:
    app: flask-monitoring
spec:
  replicas: 1
  selector:
    matchLabels:
      app: flask-monitoring
  template:
    metadata:
      labels:
        app: flask-monitoring
    spec:
      # --- START OF FIX ---
      imagePullSecrets:
        - name: ecr-read-secret
      # --- END OF FIX ---
      containers:
        - name: my-flask-app
          # IMPORTANT: Use the image name you successfully built
          image: "982068586284.dkr.ecr.us-east-1.amazonaws.com/cloud-native-monitoring-app:latest"
          ports:
            - containerPort: 5000

          # 🟢 READINESS PROBE: Is the app ready to serve traffic? (Uses /readyz)
          # This prevents the Service from routing traffic to a Pod that is still initializing.
          readinessProbe:
            httpGet:
              path: /readyz
              port: 5000
            initialDelaySeconds: 15 # Give Gunicorn workers time to spin up
            periodSeconds: 5 # Check frequently
            failureThreshold: 5 # If it fails 5 times, it's considered unready

          # 🚨 LIVENESS PROBE: Is the app process alive and responding? (Uses /healthz)
          # If this fails, Kubernetes restarts the container, enabling self-healing.
          livenessProbe:
            httpGet:
              path: /healthz
              port: 5000
            initialDelaySeconds: 20 # Start checking after the app should definitely be ready
            periodSeconds: 10 # Check every 10 seconds
            timeoutSeconds: 3
          # --- Resource Limits (Best practice for EKS) ---
          resources:
            # Both 'limits' and 'requests' must be indented one level under 'resources'
            limits:
              memory: "128Mi"
              cpu: "250m"
            requests:
              memory: "64Mi"
              cpu: "100m"