Predictive Configuration

Beyond specifying models, other configuration options can be set in the predictiveConfig YAML:

Providing Predictive Configuration

Predictive Configuration is provided through environment variables, which can be supplied through the Custom Pod Autoscaler YAML shorthand:

apiVersion: custompodautoscaler.com/v1
kind: CustomPodAutoscaler
metadata:
  name: simple-linear-example
spec:
  template:
    spec:
      containers:
      - name: simple-linear-example
        image: jthomperoo/predictive-horizontal-pod-autoscaler:latest
        imagePullPolicy: Always
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: php-apache
  roleRequiresMetricsServer: true
  config:
    - name: minReplicas
      value: "1"
    - name: maxReplicas
      value: "5"
    - name: predictiveConfig
      value: |
        models:
        - type: Linear
          name: LinearPrediction
          perInterval: 1
          linear:
            lookAhead: 10000
            storedValues: 6
        decisionType: "maximum"
        metrics:
        - type: Resource
          resource:
            name: cpu
            target:
              averageUtilization: 50
              type: Utilization
    - name: interval
      value: "10000"

The predictiveConfig is provided through this environment variable, and represented in YAML:

    - name: predictiveConfig
      value: |
        models:
        - type: Linear
          name: LinearPrediction
          perInterval: 1
          linear:
            lookAhead: 10000
            storedValues: 6
        decisionType: "maximum"

decisionType

Example:

decisionType: mean

Default value: maximum. Possible values:

• maximum - pick the highest evaluation of the models.
• minimum - pick the lowest evaluation of the models.
• mean - calculate the mean number of replicas between the models.
• median - calculate the median number of replicas between the models.

Decider on which evaluation to pick if there are multiple models provided.

dbPath

Example:

dbPath: "/tmp/path/store.db"

Default value: /store/predictive-horizontal-pod-autoscaler.db.

The path to store the SQLite3 database, e.g. for storing the DB in a volume to persist it.

migrationPath

Example:

migrationPath: "/tmp/migrations/sql"

Default value: /app/sql.

The path of the SQL migrations for the SQLite3 database.

models

List of statistical models to apply. See the models section for details.

metrics

List of metrics to target for evaluating replica counts. See the metrics section for details.

cpuInitializationPeriod

Example:

cpuInitializationPeriod: 150

Default value: 300 (5 minutes). Set in seconds. Equivalent to --horizontal-pod-autoscaler-cpu-initialization-period; the period after pod start when CPU samples might be skipped.

initialReadinessDelay

Example:

initialReadinessDelay: 45

Default value: 30 (30 seconds). Set in seconds. Equivalent to --horizontal-pod-autoscaler-initial-readiness-delay; the period after pod start during which readiness changes will be treated as initial readiness.

tolerance

Example:

tolerance: 0.25

Default value: 0.1. Equivalent to --horizontal-pod-autoscaler-tolerance; the minimum change (from 1.0) in the desired-to-actual metrics ratio for the horizontal pod autoscaler to consider scaling.