14-startup-boost.yaml

---
# Startup Boost: temporary CPU boost for cold-start workloads.
#
# Some applications (JVMs, ML model loaders, interpreted runtimes) need
# significantly more CPU during startup than at steady state. Rather than
# over-provisioning permanently, startupBoost applies a temporary CPU
# multiplier that expires after a configurable duration.
#
# How it works:
#   1. When the operator detects a new pod (within the boost duration
#      window), it resizes the pod's CPU request to
#      `recommended_cpu * multiplier`.
#   2. Once the duration expires, the operator resizes back to the
#      steady-state recommendation.
#   3. If the boosted CPU would exceed the container's CPU limit or the
#      node's allocatable CPU, the boost is capped automatically.
#
# Prerequisites:
#   - Kubernetes 1.32+ with In-Place Pod Resize enabled (1.32 requires InPlacePodVerticalScaling feature gate)
#   - Prometheus reachable at the address below
#   - attune operator installed
#
# Apply:
#   kubectl apply -f examples/14-startup-boost.yaml
apiVersion: attune.io/v1alpha1
kind: AttunePolicy
metadata:
  name: jvm-api
  namespace: production
spec:
  targetRef:
    kind: Deployment
    name: jvm-api
  metricsSource:
    prometheus:
      address: http://prometheus-server.monitoring:80
  cpu:
    percentile: 95
    overhead: "20"
    startupBoost:
      multiplier: "3.0"   # 3x the steady-state recommendation during startup
      duration: 2m         # boost expires 2 minutes after pod creation
    minAllowed: "100m"
    maxAllowed: "8000m"
  memory:
    percentile: 99
    overhead: "30"
    minAllowed: "256Mi"
    maxAllowed: "16Gi"
  updateStrategy:
    type: Auto
    cooldown: 1h