Add nil checks for hpa object target type values

[Lukasz-AWS]

What type of PR is this?

/kind bug

What this PR does / why we need it:

Fixes a bug where if the corresponding value (value, averageValue) of the the type of target (Value, AverageValue) is not set but the opposite type is, HPA object would pass validation but then cause kube-controller-manager to crash loop due to nil pointer dereference looking for that corresponding value. This PR adds a nil check to make sure kube-controller-manager doesn't crash. The change to fix the validation is being worked on here #120373

Which issue(s) this PR fixes:

N/A
Fixes #

Special notes for your reviewer:

Observed in 1.27 and 1.28, possibly might be an issue in earlier versions as well.
Can be replicated by deploying an hpa with a valid scaleTargetRef and adding this object to the metrics. On sync kcm will crash due to nil pointer dereference.

...
- apiVersion: autoscaling/v2
  kind: HorizontalPodAutoscaler
  spec:
    maxReplicas: 10
    metrics:
    - type: Object
      object:
        metric:
          name: requests-per-second
        describedObject:
          apiVersion: networking.k8s.io/v1beta1
          kind: Ingress
          name: main-route
        target:
          type: AverageValue
          value: 1
...

Does this PR introduce a user-facing change?

NONE

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

[k8s-ci-robot]

This issue is currently awaiting triage.

If a SIG or subproject determines this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[k8s-ci-robot]

Hi @Lukasz-AWS. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[liggitt]

/lgtm
/approve
/ok-to-test

[k8s-ci-robot]

LGTM label has been added.

Git tree hash: 308ca288672b5a830a7059e8d77174e05e35ff14

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: liggitt, Lukasz-AWS

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• pkg/controller/podautoscaler/OWNERS [liggitt]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[Lukasz-AWS]

/retest

[Lukasz-AWS]

/retest

[Lukasz-AWS]

Not cherry picking to 1.26, as 1.26 will forcibly change the type.
i.e. if I apply

      target:
        type: AverageValue
        value: "1"

it becomes with no crashes

      target:
        type: Value
        value: "1"

[liggitt]

Not cherry picking to 1.26, as 1.26 will forcibly change the type.

wait, what?

is that true when creating a brand new object?

[Lukasz-AWS]

yes appears so, this is how I tested against the other versions which led to no changes and kcm crashing:

dev-dsk-luukasz-2b-3776d48b % kubectl version --short
...
Server Version: v1.26.9-eks-f8587cb
...

dev-dsk-luukasz-2b-3776d48b % kubectl get hpa
No resources found in default namespace.

dev-dsk-luukasz-2b-3776d48b % cat hpa.yaml
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: test-126
  namespace: default
spec:
  maxReplicas: 10
  metrics:
  - object:
      describedObject:
        kind: Pod
        name: '*'
      metric:
        name: correlator_queue
      target:
        type: AverageValue
        value: "1"
    type: Object
  minReplicas: 1
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: php-apache

dev-dsk-luukasz-2b-3776d48b % kubectl apply -f hpa.yaml
horizontalpodautoscaler.autoscaling/test-126 created

dev-dsk-luukasz-2b-3776d48b % kubectl get hpa test-126 -oyaml
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  annotations:
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"autoscaling/v2","kind":"HorizontalPodAutoscaler","metadata":{"annotations":{},"name":"test-126","namespace":"default"},"spec":{"maxReplicas":10,"metrics":[{"object":{"describedObject":{"kind":"Pod","name":"*"},"metric":{"name":"correlator_queue"},"target":{"type":"AverageValue","value":"1"}},"type":"Object"}],"minReplicas":1,"scaleTargetRef":{"apiVersion":"apps/v1","kind":"Deployment","name":"php-apache"}}}
  creationTimestamp: "2023-10-06T21:19:18Z"
  name: test-126
  namespace: default
  resourceVersion: "254409"
  uid: 933c8c32-ea4a-40be-b0af-99c91d98c26a
spec:
  maxReplicas: 10
  metrics:
  - object:
      describedObject:
        kind: Pod
        name: '*'
      metric:
        name: correlator_queue
      target:
        type: Value
        value: "1"
    type: Object
  minReplicas: 1
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: php-apache
status:
  currentMetrics: null
  desiredReplicas: 0

[liggitt]

oh.... I bet that's because of v1 conversion logic, which infers the type from which *value field is non-nil (because there's no dedicated type field in v1):

kubernetes/pkg/apis/autoscaling/v1/conversion.go

Lines 49 to 54 in 755644a

	var metricType autoscaling.MetricTargetType
	if value == nil {
	metricType = autoscaling.AverageValueMetricType
	} else {
	metricType = autoscaling.ValueMetricType
	}

kubernetes/pkg/apis/autoscaling/v1/conversion.go

Lines 85 to 89 in 755644a

	if in.AverageValue == nil {
	metricType = autoscaling.ValueMetricType
	} else {
	metricType = autoscaling.AverageValueMetricType
	}

kubernetes/pkg/apis/autoscaling/v1/conversion.go

Lines 217 to 222 in 755644a

	var metricType autoscaling.MetricTargetType
	if utilization == nil {
	metricType = autoscaling.AverageValueMetricType
	} else {
	metricType = autoscaling.UtilizationMetricType
	}

prior to #114358 which was included in 1.27, HPA objects were persisted in etcd in v1 format, so every HPA object would interact with the v1 conversion

[liggitt]

go ahead and backport the change to 1.26 so the 1.26 controller is resilient to data it could encounter speaking to a 1.27 server