Kubernetes HPA Generator - Horizontal Pod Autoscaler

Generate Kubernetes HPA manifests for automatic scaling. Configure CPU, memory, and custom metrics targets with min/max replicas and scale-down behavior.

HPA Name

Namespace

API Version

Scale Target

Target Kind

Target Name

Min Replicas

Max Replicas

Metrics

CPU Utilization

Target (%)

Memory Utilization

Custom Metric

Scaling Behavior (v2)

Custom scaling behavior

HPA YAML

kubectl Commands

# Apply HPA
kubectl apply -f hpa.yaml

# Check HPA status
kubectl get hpa my-app-hpa -n default

# Watch scaling events
kubectl get hpa -w

# Describe for details
kubectl describe hpa my-app-hpa -n default

# Check metrics-server is running
kubectl get deployment metrics-server -n kube-system

Prerequisites

# Install metrics-server (required for CPU/memory metrics)
kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

# Verify metrics are available
kubectl top pods -n default

# Target deployment must have resource requests defined
# for percentage-based utilization metrics to work

Kubernetes Service Generator - Create Se…

Kubernetes CronJob Generator - Scheduled…

Kubernetes Ingress Generator - Create In…

Kubernetes ConfigMap Generator - Create …

View all Kubernetes configurators →

Frequently Asked Questions

Horizontal Pod Autoscaler automatically scales pod replicas based on metrics like CPU, memory, or custom metrics. It monitors targets and adjusts replica count within min/max bounds.

Built-in: CPU and memory utilization. Custom metrics: any metric from metrics-server or Prometheus adapter. External metrics: metrics from outside the cluster.

HPA compares current CPU usage against target (e.g., 80%). If usage exceeds target, it scales up. Formula: desiredReplicas = currentReplicas * (currentMetric / targetMetric).

Utilization is percentage of resource request (80% of 100m = 80m). AverageValue is absolute value (200m average). Use Utilization for CPU, either for memory.

Stabilization prevents flapping by considering metrics over a time window before scaling. Default: 300s for scale-down, 0s for scale-up. Prevents rapid scale changes.

Use behavior.scaleDown settings: stabilizationWindowSeconds (wait time), policies (percent or pod limits per period). Example: max 10% reduction every 60 seconds.

Standard HPA requires minReplicas >= 1. For scale-to-zero, use KEDA (Kubernetes Event-driven Autoscaling) which extends HPA with event-based triggers.

Metrics-server provides resource metrics (CPU, memory) for HPA and kubectl top. It's required for HPA to work. Install via helm or kubectl apply.

Deploy a custom metrics adapter (Prometheus Adapter, Datadog). Register metrics with Kubernetes API. Reference them in HPA spec.metrics with type: Pods or Object.

HPA scales horizontally (more pods). VPA scales vertically (bigger pods). Use HPA for stateless workloads, VPA for stateful or when pods need more resources.

Related Tools