node.yaml

apiVersion: machinelearning.seldon.io/v1
kind: SeldonDeployment
metadata:
  name: queue
spec:
  protocol: v2
  name: queue
  predictors:
  - name: queue
    annotations:
      seldon.io/no-engine: "true" 
    componentSpecs:
    - spec:
        containers:
        - image: sdghafouri/queue:queue
          name: queue
          imagePullPolicy: Always
          resources:
            requests:
              cpu: '4'
              memory: 8Gi
            limits:
              cpu: '16'
              memory: 16Gi
          env:
            - name: MODEL_NAME
              value: 'node-one'
            - name: MLSERVER_MODEL_MAX_BATCH_SIZE
              value: "5"
            - name: MLSERVER_MODEL_MAX_BATCH_TIME
              value: "1"
            - name: MLSERVER_PARALLEL_WORKERS
              value: "1"
            - name: LAST_NODE
              value: 'True'
            - name: DROP_LIMIT
              value: "5"
          readinessProbe:
            failureThreshold: 3
            initialDelaySeconds: 0
            periodSeconds: 1
            successThreshold: 1
            tcpSocket:
              port: 9000
            timeoutSeconds: 1
      replicas: 1
    graph:
      name: queue
      type: MODEL
      children: []
    labels:
      sidecar.istio.io/inject: "true"