TimeSeriesPrediction status PredictPartial message is not all metric predicted

[wxuedong]

Describe the feature

K8s version 1.20.0
Craned version 0.11.0

Create ehpa yaml:

apiVersion: autoscaling.crane.io/v1alpha1
kind: EffectiveHorizontalPodAutoscaler
metadata:
  name: biz-service
  namespace: ptest
spec:
  scaleTargetRef:
    apiVersion: argoproj.io/v1alpha1
    kind: Rollout
    name: biz-service
  minReplicas: 1        
  maxReplicas: 18      
  scaleStrategy: Auto    
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 50
  prediction:
    predictionWindowSeconds: 3600   
    predictionAlgorithm:
      algorithmType: dsp
      dsp:
        sampleInterval: "60s"
        historyLength: "7d"

kubectl describe ehpa biz-service

Annotations:  autoscaling.crane.io/effective-hpa-current-metrics:
                - resource:
                    current:
                      averageUtilization: 12
                      averageValue: 516m
                    name: cpu
                  type: Resource
...
Status:
  Conditions:
    Last Transition Time:  2024-01-11T01:34:34Z
    Message:               Effective HPA is ready
    Reason:                EffectiveHorizontalPodAutoscalerReady
    Status:                True
    Type:                  Ready
    Last Transition Time:  2024-01-11T01:34:34Z
    Message:               not all metric predicted
    Reason:                PredictPartial
    Status:                False
    Type:                  PredictionReady
    Last Transition Time:  2024-01-11T01:34:34Z
    Message:               recommended size matches current size
    Reason:                ReadyForNewScale
    Status:                True
    Type:                  AbleToScale
    Last Transition Time:  2024-01-11T01:34:34Z
    Message:               the HPA was able to successfully calculate a replica count from cpu resource utilization (percentage of request)
    Reason:                ValidMetricFound
    Status:                True
    Type:                  ScalingActive
    Last Transition Time:  2024-01-11T01:34:34Z
    Message:               the desired replica count is less than the minimum replica count
    Reason:                TooFewReplicas
    Status:                True
    Type:                  ScalingLimited
  Current Replicas:        16
  Expect Replicas:         16
Events:
  Type    Reason             Age    From                      Message
  ----    ------             ----   ----                      -------
  Normal  PredictionCreated  6m22s  effective-hpa-controller  Create TimeSeriesPrediction successfully
  Normal  HPACreated         6m22s  effective-hpa-controller  Create HorizontalPodAutoscaler successfully

kubectl describe hpa ehpa-biz-service

Metrics:                                               ( current / target )
  resource cpu on pods  (as a percentage of request):  12% (517m) / 50%
...

kubectl describe TimeSeriesPrediction ehpa-biz-service

Spec:
  Prediction Metrics:
    Algorithm:
      Algorithm Type:  dsp
      Dsp:
        Estimators:
        History Length:   2d
        Sample Interval:  60s
    Expression Query:
      Expression:             sum(irate(container_cpu_usage_seconds_total{namespace="ptest",pod=~"^biz-service-.*",container!=""}[3m]))
    Resource Identifier:      resource.cpu
    Type:                     ExpressionQuery
  Prediction Window Seconds:  3600
  Target Ref:
    API Version:  argoproj.io/v1alpha1
    Kind:         Rollout
    Name:         biz-service
    Namespace:    ptest
Status:
  Conditions:
    Last Transition Time:  2024-01-11T01:31:33Z
    Message:               not all metric predicted
    Reason:                PredictPartial
    Status:                False
    Type:                  Ready
  Prediction Metrics:
    Ready:                false
    Resource Identifier:  resource.cpu

prometheus data:

ps: biz-service has been running for more than three months

Why is there no crane_autoscaling_prediction Metric in HPA resources ?

Under what circumstances will there be the message "not all metric predicted"?

[yangyegang]

I have the same problem, have you solved it?