Scaling Applications

Scaling Applications

Scaling in Kubernetes means adjusting the number of replicas for a workload. You can scale manually (changing the replica count) or automatically (using the Horizontal Pod Autoscaler).

Manual Scaling

# Scale a deployment
kubectl scale deployment my-app --replicas=5

# Scale from a manifest
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app
spec:
  replicas: 5     # Change this number
  # ... rest of spec

Manual scaling is simple and predictable. Use it when you know exactly how many replicas you need.

Understanding Resource Requests

containers:
  - name: app
    image: my-app
    resources:
      requests:
        cpu: "250m"       # 0.25 CPU cores
        memory: "128Mi"   # 128 MB RAM
      limits:
        cpu: "500m"       # 0.5 CPU cores
        memory: "256Mi"   # 256 MB RAM

requests — The minimum resources a pod needs. The scheduler uses this to decide which node to place the pod on.

limits — The maximum resources a pod can use. If it exceeds this, it may be throttled or killed.

Scaling Best Practices

Always set resource requests — Without them, the scheduler cannot make informed placement decisions.

Start small, scale up — Begin with 2-3 replicas and increase based on load.

Use Pod Disruption Budgets — Ensure a minimum number of pods are always available during scaling events.

apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
  name: my-app-pdb
spec:
  minAvailable: 2    # At least 2 pods must always be running
  selector:
    matchLabels:
      app: my-app

Scaling with kubectl

# Scale up
kubectl scale deployment my-app --replicas=10

# Verify
kubectl get deployment my-app
kubectl get pods -l app=my-app

# Autoscale (basic — sets min, max, and target CPU)
kubectl autoscale deployment my-app --min=2 --max=10 --cpu-percent=70

The kubectl autoscale command creates a Horizontal Pod Autoscaler, which we will cover in the advanced section.