Support for reporting authorization errors during scale up

cluster-autoscaler/core/scale_up.go

@@ -666,9 +666,13 @@ func executeScaleUp(context *context.AutoscalingContext, clusterStateRegistry *c
 	increase := info.NewSize - info.CurrentSize
 	if err := info.Group.IncreaseSize(increase); err != nil {
 		context.LogRecorder.Eventf(apiv1.EventTypeWarning, "FailedToScaleUpGroup", "Scale-up failed for group %s: %v", info.Group.Id(), err)
-		clusterStateRegistry.RegisterFailedScaleUp(info.Group, metrics.CloudProviderError, now)
+		reason := metrics.CloudProviderError
-		return errors.NewAutoscalerError(errors.CloudProviderError,
+		aerr := errors.ToAutoscalerError(errors.CloudProviderError, err).AddPrefix("failed to increase node group size: %v", err)
-			"failed to increase node group size: %v", err)
+		if aerr.Type() == errors.AuthorizationError {
+			reason = metrics.AuthorizationError
+		}
+		clusterStateRegistry.RegisterFailedScaleUp(info.Group, reason, now)
+		return aerr
 	}
 	clusterStateRegistry.RegisterOrUpdateScaleUp(
 		info.Group,

cluster-autoscaler/core/scale_up_test.go

@@ -23,13 +23,15 @@ import (
 	"testing"
 	"time"
 
-	"k8s.io/autoscaler/cluster-autoscaler/core/utils"
-
 	"k8s.io/autoscaler/cluster-autoscaler/cloudprovider"
+	mockprovider "k8s.io/autoscaler/cluster-autoscaler/cloudprovider/mocks"
 	testprovider "k8s.io/autoscaler/cluster-autoscaler/cloudprovider/test"
 	"k8s.io/autoscaler/cluster-autoscaler/clusterstate"
 	"k8s.io/autoscaler/cluster-autoscaler/config"
+	"k8s.io/autoscaler/cluster-autoscaler/core/utils"
 	"k8s.io/autoscaler/cluster-autoscaler/estimator"
+	"k8s.io/autoscaler/cluster-autoscaler/processors/nodegroupset"
+	"k8s.io/autoscaler/cluster-autoscaler/utils/errors"
 	kube_util "k8s.io/autoscaler/cluster-autoscaler/utils/kubernetes"
 	. "k8s.io/autoscaler/cluster-autoscaler/utils/test"
 	"k8s.io/autoscaler/cluster-autoscaler/utils/units"
@@ -970,3 +972,19 @@ func TestCheckScaleUpDeltaWithinLimits(t *testing.T) {
 		}
 	}
 }
+
+func TestAuthError(t *testing.T) {
+	context, err := NewScaleTestAutoscalingContext(config.AutoscalingOptions{}, &fake.Clientset{}, nil, nil, nil)
+	assert.NoError(t, err)
+
+	nodeGroup := &mockprovider.NodeGroup{}
+	info := nodegroupset.ScaleUpInfo{Group: nodeGroup}
+	nodeGroup.On("Id").Return("A")
+	nodeGroup.On("IncreaseSize", 0).Return(errors.NewAutoscalerError(errors.AuthorizationError, ""))
+
+	clusterStateRegistry := clusterstate.NewClusterStateRegistry(nil, clusterstate.ClusterStateRegistryConfig{}, context.LogRecorder, newBackoff())
+
+	aerr := executeScaleUp(&context, clusterStateRegistry, info, "", time.Now())
+	assert.Error(t, aerr)
+	assert.Equal(t, errors.AuthorizationError, aerr.Type())
+}

cluster-autoscaler/metrics/metrics.go

@@ -65,6 +65,8 @@ const (
 	APIError FailedScaleUpReason = "apiCallError"
 	// Timeout was encountered when trying to scale-up
 	Timeout FailedScaleUpReason = "timeout"
+	// AuthorizationError is an authorization error.
+	AuthorizationError FailedScaleUpReason = "authorizationError"
 
 	// autoscaledGroup is managed by CA
 	autoscaledGroup NodeGroupType = "autoscaled"

cluster-autoscaler/utils/errors/errors.go

@@ -61,6 +61,8 @@ const (
 	// NodeGroupDoesNotExistError signifies that a NodeGroup
 	// does not exist.
 	NodeGroupDoesNotExistError AutoscalerErrorType = "nodeGroupDoesNotExistError"
+	// AuthorizationError signifies that an authorization error occurred.
+	AuthorizationError AutoscalerErrorType = "authorizationError"
 )
 
 // NewAutoscalerError returns new autoscaler error with a message constructed from format string