Kubernetes Autoscaler Threshold Calculator

The Horizontal Pod Autoscaler (HPA) automatically scales Deployment replicas based on CPU, memory, or custom metrics. Correct threshold configuration is critical — wrong values cause either constant scale-flapping or under-utilized pods.

HPA v2 Spec Example

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: my-app-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: my-app
  minReplicas: 2
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: AverageValue
        averageValue: 350m   # 70% of 500m limit
  behavior:
    scaleDown:
      stabilizationWindowSeconds: 300

Scaling Algorithm

desired_replicas = ceil(current_replicas × (current_metric / target_metric))

If 3 pods average 700m CPU with a 350m target: ceil(3 × (700/350)) = 6 replicas.

Common Mistakes

1. 1.Target on limits not requests: HPA type: Utilization targets a % of _requests_, not limits. If your requests are lower than limits, you'll scale out sooner than expected.
2. 2.No stabilization window: Default scale-down is immediate — configure a stabilization window to prevent yo-yo scaling.
3. 3.CPU limits too low: If your app regularly hits CPU limits (throttling), it will never trigger HPA scale-out — it just runs slow instead.
4. 4.Max replicas too low: If maxReplicas is too small, HPA can't scale to meet demand. Monitor kube_horizontalpodautoscaler_status_current_replicas == kube_horizontalpodautoscaler_spec_max_replicas for this condition.