{
  "description": "Example SeldonDeployment with a canary deployment strategy splitting traffic 90/10 between stable and canary predictors",
  "request": {
    "method": "POST",
    "url": "http://seldon-deploy/seldon-deploy/api/v1alpha1/namespaces/production/seldondeployments",
    "headers": {
      "Content-Type": "application/json",
      "Authorization": "Bearer <oidc-token>"
    },
    "body": {
      "apiVersion": "machinelearning.seldon.io/v1",
      "kind": "SeldonDeployment",
      "metadata": {
        "name": "iris-canary",
        "namespace": "production"
      },
      "spec": {
        "name": "iris-canary",
        "predictors": [
          {
            "name": "stable",
            "traffic": 90,
            "replicas": 2,
            "graph": {
              "name": "classifier",
              "type": "MODEL",
              "implementation": "SKLEARN_SERVER",
              "modelUri": "gs://seldon-models/iris/v1"
            }
          },
          {
            "name": "canary",
            "traffic": 10,
            "replicas": 1,
            "graph": {
              "name": "classifier",
              "type": "MODEL",
              "implementation": "SKLEARN_SERVER",
              "modelUri": "gs://seldon-models/iris/v2"
            }
          }
        ]
      }
    }
  },
  "response": {
    "status": 201,
    "body": {
      "apiVersion": "machinelearning.seldon.io/v1",
      "kind": "SeldonDeployment",
      "metadata": {
        "name": "iris-canary",
        "namespace": "production"
      },
      "status": {
        "state": "Creating"
      }
    }
  }
}