# services/bstein-dev-home/backend-deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: bstein-dev-home-backend
  namespace: bstein-dev-home
spec:
  replicas: 1
  revisionHistoryLimit: 3
  selector:
    matchLabels:
      app: bstein-dev-home-backend
  template:
    metadata:
      labels:
        app: bstein-dev-home-backend
    spec:
      automountServiceAccountToken: true
      serviceAccountName: bstein-dev-home
      nodeSelector:
        kubernetes.io/arch: arm64
        node-role.kubernetes.io/worker: "true"
      imagePullSecrets:
        - name: harbor-bstein-robot
      containers:
        - name: backend
          image: registry.bstein.dev/bstein/bstein-dev-home-backend:0.1.1-23 # {"$imagepolicy": "bstein-dev-home:bstein-dev-home-backend"}
          imagePullPolicy: Always
          env:
            - name: AI_CHAT_API
              value: http://ollama.ai.svc.cluster.local:11434
            - name: AI_CHAT_MODEL
              value: qwen2.5-coder:7b-instruct-q4_0
            - name: AI_CHAT_TIMEOUT_SEC
              value: "20"
            - name: AI_NODE_NAME
              valueFrom:
                fieldRef:
                  fieldPath: spec.nodeName
            - name: AI_NODE_GPU_MAP
              value: |
                {"titan-24": "RTX 3080 8GB (local GPU)", "titan-22": "RTX 3050 8GB (local GPU)"}
          ports:
            - name: http
              containerPort: 8080
          readinessProbe:
            httpGet:
              path: /api/healthz
              port: http
            initialDelaySeconds: 2
            periodSeconds: 5
          livenessProbe:
            httpGet:
              path: /api/healthz
              port: http
            initialDelaySeconds: 10
            periodSeconds: 10
          resources:
            requests:
              cpu: 50m
              memory: 64Mi
            limits:
              cpu: 300m
              memory: 256Mi