Dry run deployment config

[zakazai]

Context

I'm trying to validate a deployment config at server

$ cat deploy_test.yaml
apiVersion: machinelearning.seldon.io/v1
kind: SeldonDeployment
metadata:
  name: iris-model-test
  namespace: seldon-service
spec:
  name: iris
  predictors:
  - graph:
      implementation: SKLEARN_SERVER
      modelUri: gs://seldon-models/sklearn/iris
      name: classifier
    name: default
    replicas: 1

and got this result

$ kubectl apply -f deploy_test.yaml --dry-run=server
Error from server (BadRequest): error when applying patch:
{"spec":{"predictors":[{"graph":{"implementation":"SKLEARN_SERVER","modelUri":"gs://seldon-models/sklearn/iris","name":"classifier"},"name":"default","replicas":1}]}}
to:
Resource: "machinelearning.seldon.io/v1, Resource=seldondeployments", GroupVersionKind: "machinelearning.seldon.io/v1, Kind=SeldonDeployment"
Name: "iris-model-test", Namespace: "seldon-service"
for: "deploy_test.yaml": admission webhook "v1.mseldondeployment.kb.io" does not support dry run

Currently I'm trying to create a CI test to validate my deployment config during pull request.
I'm not sure if this already possible or not.

Environments

• minikube version: v1.15.0
• seldon-core-operator-1.4.0
• kubectl Client Version "v1.19.4", Server Version "v1.19.4"

[ukclivecox]

I think we are missing the sideEffects directive in the Webhook definitions. As its not there defaut to UnKnown and therefore dry run is rejected.
https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.19/#mutatingwebhook-v1-admissionregistration-k8s-io

I think as we don't have sideEffects this could be changed.