Optimize Kubernetes Autoscaling Configuration

This skill provides comprehensive expertise in Kubernetes autoscaling technologies: Horizontal Pod Autoscaler (HPA), Vertical Pod Autoscaler (VPA), and Cluster Autoscaler. It delivers production-ready configuration guidance, optimization strategies, and best practices for implementing efficient autoscaling in Kubernetes environments. The skill covers resource-based scaling, custom metrics integration, stabilization policies, and multi-tier autoscaling architectures.

Use this skill when configuring autoscaling for production workloads, optimizing resource utilization across pod and cluster levels, implementing custom metrics-based scaling for queue workers or event-driven applications, or troubleshooting scaling behavior and flapping issues. Use it when designing multi-tier autoscaling strategies that combine HPA for immediate traffic response, VPA for long-term resource optimization, and Cluster Autoscaler for node capacity management.

Do NOT use this skill for static workloads with predictable resource requirements where manual scaling is sufficient. Avoid using VPA in Auto mode for stateful applications like databases where pod restarts could cause service disruption - use recommendation mode instead.

You provide details about your workload type (web application, queue worker, database, batch job), current resource usage patterns, scaling requirements (min/max replicas, target utilization), and metrics you want to scale on (CPU, memory, custom metrics like queue depth). You receive production-ready YAML configurations for HPA, VPA, and Cluster Autoscaler, complete with stabilization windows, scaling policies, resource boundaries, and monitoring setup.

Here's a complete HPA configuration example with CPU and memory metrics:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: web-app-hpa
  namespace: production
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: web-app
  minReplicas: 3
  maxReplicas: 100
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70
  - type: Resource
    resource:
      name: memory
      target:
        type: Utilization
        averageUtilization: 80
  behavior:
    scaleDown:
      stabilizationWindowSeconds: 300
      policies:
      - type: Percent
        value: 50
        periodSeconds: 60
      - type: Pods
        value: 2
        periodSeconds: 60
      selectPolicy: Min
    scaleUp:
      stabilizationWindowSeconds: 60
      policies:
      - type: Percent
        value: 100
        periodSeconds: 15
      - type: Pods
        value: 4
        periodSeconds: 15
      selectPolicy: Max

The skill covers configuration for GKE (Google Kubernetes Engine) with Cluster Autoscaler node pool auto-discovery, Prometheus with ServiceMonitor for metrics collection and alerting, kube-state-metrics for HPA metrics exposure, custom metrics adapters for external metrics like SQS queue length, and metrics server for resource-based autoscaling. It includes PodDisruptionBudgets to maintain availability during scaling operations.

Kubernetes platform engineers designing autoscaling strategies for production clusters, DevOps teams optimizing resource utilization and cost efficiency, SREs troubleshooting scaling behavior and preventing flapping, and application developers implementing custom metrics-based autoscaling for microservices, queue workers, or event-driven architectures. The skill provides pattern-specific guidance: CPU-based HPA with 70-80% target utilization for web applications, custom metrics HPA based on queue depth for queue workers, VPA in recommendation mode with manual tuning for databases, and Cluster Autoscaler with job-specific node pools for batch jobs.