enpoint config `endpoint.httpPort` in graph does'n take effect.

[cnfatal]

Describe the bug

Configuration httpPort and grpcPort in .spec.predictors[].graph.endpoint do not take effect.

To reproduce

apiVersion: machinelearning.seldon.io/v1
kind: SeldonDeployment
metadata:
  name: sklearn-iris
spec:
  predictors:
    - name: default
      replicas: 1
      annotations:
        seldon.io/no-engine: "true"
      graph:
        name: classifier
        endpoint:
          type: REST
          httpPort: 80  # <--- here
      componentSpecs:
        - spec:
            containers:
              - name: classifier
                image: nginx:1.13
                securityContext:
                  privileged: true

$ kubectl -n seldon get svc sklearn-iris-default -oyaml
apiVersion: v1
kind: Service
metadata:
  ...
  ports:
  - name: http
    port: 9000
    protocol: TCP
    targetPort: 9000
  - name: grpc
    port: 9500
    protocol: TCP
    targetPort: 9500
  selector:
    seldon-app: sklearn-iris-default
...

But it works if :

apiVersion: machinelearning.seldon.io/v1
kind: SeldonDeployment
metadata:
  name: sklearn-iris
spec:
  predictors:
    - name: default
      replicas: 1
      annotations:
        seldon.io/no-engine: "true"
      graph:
        name: classifier
        # endpoint:
        #   type: REST
        #   httpPort: 80
      componentSpecs:
        - spec:
            containers:
              - name: classifier
                image: nginx:1.13
                securityContext:
                  privileged: true
                ports:
                  - containerPort: 80
                    protocol: TCP
                    name: http

$ kubectl -n seldon get svc sklearn-iris-default -oyaml
apiVersion: v1
kind: Service
...
  ports:
  - name: http
    port: 80
    protocol: TCP
    targetPort: 80
  - name: grpc
    port: 9500
    protocol: TCP
    targetPort: 9500
  selector:
    seldon-app: sklearn-iris-default
...

Expected behaviour

Environment

Model Details

• Images of your model: [Output of: kubectl get seldondeployment -n <yourmodelnamespace> <seldondepname> -o yaml | grep image: where <yourmodelnamespace>]
• Logs of your model: [You can get the logs of your model by running kubectl logs -n <yourmodelnamespace> <seldonpodname> <container>]

[ukclivecox]

This is the code that needs to be looked at

[ukclivecox]

Yes at present you need to override this in the container spec. Will close this for now. If others need more configurability please add to the issue.