apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: api-hpa
spec:
  minReplicas: 3
  maxReplicas: 30
  behavior:
    scaleUp:
      stabilizationWindowSeconds: 60
      policies:
      - type: Percent
        value: 50
        periodSeconds: 60
    scaleDown:
      stabilizationWindowSeconds: 300
      policies:
      - type: Percent
        value: 20
        periodSeconds: 60
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: api
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 60
  - type: Pods
    pods:
      metric:
        name: request_rate
      target:
        type: AverageValue
        averageValue: "50"   # pod당 RPS 타깃

apiVersion: keda.sh/v1alpha1
kind: ScaledObject
metadata:
  name: worker-queue
spec:
  scaleTargetRef:
    name: worker
  minReplicaCount: 2
  maxReplicaCount: 50
  cooldownPeriod: 300
  triggers:
  - type: rabbitmq
    metadata:
      protocol: amqp
      queueName: jobs
      hostFromEnv: RABBITMQ_CONN
      queueLength: "200"   # 포드당 200개 목표

upstream api {
  least_conn;
  server 10.0.0.11:8080 max_fails=3 fail_timeout=10s;
  server 10.0.0.12:8080 max_fails=3 fail_timeout=10s;
}
map $upstream_connect_time $is_warm { default 1; ~^0\..* 0; }
server {
  listen 443 ssl http2;
  location / {
    proxy_set_header X-Warm $is_warm;
    proxy_next_upstream error timeout http_502 http_503;
    proxy_pass http://api;
  }
}