apiVersion: apps/v1
kind: Deployment
metadata:
  name: ai-service-worker
  namespace: ai-service
spec:
  replicas: 1
  selector:
    matchLabels:
      app: ai-service-worker
  template:
    metadata:
      labels:
        app: ai-service-worker
    spec:
      terminationGracePeriodSeconds: 20
      hostAliases:
        - ip: "77.105.173.42"
          hostnames:
            - "s3-minio.estateliga.work"
      containers:
        - name: worker
          image: localhost:30300/admin/ai-service:latest
          command: ["/usr/local/bin/ai-service-worker"]
          ports:
            - containerPort: 8081
          env:
            - name: WORKER_ID
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
            - name: WORKER_TASK_TYPES
              value: "llm_chat,chat_completion,call_analysis,telegram_classification,transcript_summary"
            - name: WORKER_MODEL_PROFILES
              value: "qwen2.5-14b"
            - name: WORKER_CLAIM_LIMIT
              value: "8"
            - name: WORKER_LEASE_TIMEOUT
              value: "5m"
          envFrom:
            - configMapRef:
                name: ai-service-config
            - secretRef:
                name: ai-service-secrets
          startupProbe:
            httpGet:
              path: /readyz
              port: 8081
            periodSeconds: 5
            failureThreshold: 30
          readinessProbe:
            httpGet:
              path: /readyz
              port: 8081
            periodSeconds: 10
          livenessProbe:
            httpGet:
              path: /healthz
              port: 8081
            periodSeconds: 10
          resources:
            requests:
              cpu: 50m
              memory: 96Mi
            limits:
              cpu: 500m
              memory: 384Mi
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: ai-service-transcription-worker
  namespace: ai-service
spec:
  replicas: 1
  selector:
    matchLabels:
      app: ai-service-transcription-worker
  template:
    metadata:
      labels:
        app: ai-service-transcription-worker
    spec:
      terminationGracePeriodSeconds: 20
      hostAliases:
        - ip: "77.105.173.42"
          hostnames:
            - "s3-minio.estateliga.work"
      containers:
        - name: worker
          image: localhost:30300/admin/ai-service:latest
          command: ["/usr/local/bin/ai-service-worker"]
          ports:
            - containerPort: 8081
          env:
            - name: WORKER_ID
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
            - name: WORKER_TASK_TYPES
              value: "transcription"
            - name: WORKER_MODEL_PROFILES
              value: "whisper-large-v3"
            - name: WORKER_CLAIM_LIMIT
              value: "4"
            - name: WORKER_LEASE_TIMEOUT
              value: "5m"
          envFrom:
            - configMapRef:
                name: ai-service-config
            - secretRef:
                name: ai-service-secrets
          startupProbe:
            httpGet:
              path: /readyz
              port: 8081
            periodSeconds: 5
            failureThreshold: 30
          readinessProbe:
            httpGet:
              path: /readyz
              port: 8081
            periodSeconds: 10
          livenessProbe:
            httpGet:
              path: /healthz
              port: 8081
            periodSeconds: 10
          resources:
            requests:
              cpu: 50m
              memory: 96Mi
            limits:
              cpu: 500m
              memory: 384Mi