Automating Kubernetes Pod Scaling Issues – Best Practices & Real‑World Solutions

Target audience: software developers, engineers, and DevOps professionals.

Why Pod Scaling Fails

Even with the Horizontal Pod Autoscaler (HPA) enabled, many teams hit roadblocks such as:

Metrics‑server latency
Incorrect resource requests/limits
Scaling cooldown conflicts
Missing custom metrics

Understanding the root cause is the first step toward reliable automation.

Step‑by‑Step Troubleshooting Guide

1️⃣ Verify Metrics Server

kubectl get deployment metrics-server -n kube-system
kubectl logs deployment/metrics-server -n kube-system

If the pod is CrashLoopBackOff, reinstall:

kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

2️⃣ Check HPA Configuration

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: web-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: web
  minReplicas: 2
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 60

Make sure cpu requests are set on the target pods; otherwise HPA cannot compute utilization.

3️⃣ Add Custom Metrics (Optional)

For request‑rate‑based scaling you might use Prometheus Adapter:

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm install prometheus-adapter prometheus-community/prometheus-adapter

Then reference the metric in the HPA:

metrics:
- type: Pods
  pods:
    metric:
      name: http_requests_per_second
    target:
      type: AverageValue
      averageValue: 100

4️⃣ Automate Healing with a Bash Guard

Below is a tiny script that watches the HPA status and forces a re‑sync when it gets stuck:

#!/usr/bin/env bash
NAMESPACE=default
HPA_NAME=web-hpa
while true; do
  STATUS=$(kubectl get hpa $HPA_NAME -n $NAMESPACE -o jsonpath='{.status.conditions[?(@.type=="AbleToScale")].status}')
  if [[ "$STATUS" != "True" ]]; then
    echo "[$(date)] HPA not able to scale – restarting metrics‑server"
    kubectl rollout restart deployment/metrics-server -n kube-system
  fi
  sleep 60
done