# 06 — HorizontalPodAutoscaler: add/remove Pods based on load. Needs
# metrics-server (CPU/memory) or an adapter (custom/external metrics like
# queue depth or RPS).
#
#   kubectl apply -f 06-hpa.yaml
#   kubectl get hpa checkout --watch
#   # generate load, watch REPLICAS climb, then settle after cooldown
#
# HPA only works if the Deployment sets resource *requests* — the target
# utilisation is a percentage OF the request. No request => no scaling.
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: checkout
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: checkout
  minReplicas: 4
  maxReplicas: 40
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 65
    - type: Resource
      resource:
        name: memory
        target:
          type: Utilization
          averageUtilization: 80
  behavior:
    scaleUp:
      # React fast to traffic spikes: allow doubling every 30s.
      stabilizationWindowSeconds: 0
      policies:
        - type: Percent
          value: 100
          periodSeconds: 30
        - type: Pods
          value: 4
          periodSeconds: 30
      selectPolicy: Max
    scaleDown:
      # Scale in slowly so a brief dip doesn't thrash capacity.
      stabilizationWindowSeconds: 300
      policies:
        - type: Percent
          value: 10
          periodSeconds: 60