Note_Tech

All technological notes.

---

Project maintained by simonangel-fong Hosted on GitHub Pages — Theme by mattgraham

Kubernetes - Pod: Horizontal Scaling

Back

• Kubernetes - Pod: Horizontal Scaling
  • Manually Horizontal Scaling
    • Lab: Manual Scaling
  • Horizontal Pod Autoscaler(HPA)
    • Imperative Commands
    • Declarative method
    • Lab: Create HPA - Imperative command
    • Lab: Create HPA - Declarative method

---

Manually Horizontal Scaling

• When there are insufficient resources for cluster,
  • the command kubectl scale leads to some pods reach the limit, and unscaled pods remain pending state.

CMD	DESC
kubectl top nodes	List all nodes’ CPU and memory consumption
kubectl top pods	List all pods’ CPU and memory consumption
kubectl scale rs/NAME --replicas=N	Scale a ReplicaSet directly (rare—usually let the Deployment manage it).
kubectl scale deploy/NAME --replicas=N	Scale Deployment replica count. Preferred for stateless services.
kubectl scale sts/NAME --replicas=N	Scale a StatefulSet (ordered, one-by-one; respect storage/identity).
kubectl edit deploy/NAME	Open editor; set .spec.replicas manually, then save to apply.
kubectl patch deploy/NAME -p '{"spec":{"replicas":N}}'	Quick JSON patch to set replicas without an editor.

---

Lab: Manual Scaling

apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app
spec:
  replicas: 1
  selector:
    matchLabels:
      app: my-app
  template:
    metadata:
      labels:
        app: my-app
    spec:
      containers:
        - name: my-app
          image: nginx
          resources:
            requests:
              cpu: "250m"
            limits:
              cpu: "500m"

kubectl create -f app-deploy.yaml && kubectl rollout status deploy my-app
# deployment.apps/my-app created
# Waiting for deployment "my-app" rollout to finish: 0 of 1 updated replicas are available...
# deployment "my-app" successfully rolled out

# confirm
kubectl get pod
# NAME                      READY   STATUS    RESTARTS        AGE
# my-app-7cb77b7584-dcf5f   1/1     Running   0                20s

# manually monitor resource consumption
kubectl top pod my-app-7cb77b7584-dcf5f
# NAME                      CPU(cores)   MEMORY(bytes)
# my-app-7cb77b7584-dcf5f   0m           4Mi

# manually scale pod
kubectl scale deploy my-app --replicas=3 && kubectl rollout status deploy my-app
# deployment.apps/my-app scaled
# Waiting for deployment "my-app" rollout to finish: 1 out of 3 new replicas have been updated...
# Waiting for deployment "my-app" rollout to finish: 1 of 3 updated replicas are available...
# Waiting for deployment "my-app" rollout to finish: 2 of 3 updated replicas are available...
# deployment "my-app" successfully rolled out

# confirm
kubectl get deploy
# NAME     READY   UP-TO-DATE   AVAILABLE   AGE
# my-app   3/3     3            3           2m33s

# cleanup
kubectl delete deploy my-app
# deployment.apps "my-app" deleted from default namespace

---

Horizontal Pod Autoscaler(HPA)

• Horizontal Pod Autoscaler(HPA)

  • an API resource and controller that automatically adjusts the number of Pod replicas in a Deployment, StatefulSet, or ReplicaSet.
    • continuously monitor the metrics,
    • automatically scale the number of pods,
    • balance the thresholds,
    • track multiple metrics.
  • since version 1.23
  • prerequistie: metric server
  • metrics from
    • internal sources:
      • metrics server
      • custom metric adapter
    • external sources
      • datadog
      • dynatrace

• HPA requires

  • metric server
  • resource limits defined in pod
    • otherwise the target in hpa shows unknown instead of the % of the CPU usage.

---

Imperative Commands

• Manage

CMD	DESC
kubectl top pods	Check live CPU/MEM.
kubectl get hpa	List HPAs.
kubectl describe hpa/NAME	Show details: metrics, events, scale decisions, stabilization windows.
kubectl edit hpa/NAME	Change min/max replicas, metrics, or behavior.
kubectl delete hpa/NAME	Remove autoscaling so manual replicas sticks.

• Create

CMD	DESC
kubectl autoscale deploy/NAME --min=MIN --max=MAX --cpu-percent=TARGET	Create an HPA for a Deployment.
kubectl autoscale sts/NAME --min=MIN --max=MAX --cpu-percent=TARGET	Create an HPA for a StatefulSet.

---

Declarative method

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

---

Lab: Create HPA - Imperative command

• Create deployment

# app-deploy.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: my-app
spec:
  replicas: 1
  selector:
    matchLabels:
      app: my-app
  template:
    metadata:
      labels:
        app: my-app
    spec:
      containers:
        - name: my-app
          image: nginx
          resources:
            requests:
              cpu: "250m"
            limits:
              cpu: "500m"

kubectl create -f app-deploy.yaml && kubectl rollout status deploy my-app
# deployment.apps/my-app created
# Waiting for deployment "my-app" rollout to finish: 0 of 1 updated replicas are available...
# deployment "my-app" successfully rolled out

# confirm
kubectl get deploy
# NAME     READY   UP-TO-DATE   AVAILABLE   AGE
# my-app   1/1     1            1           114s

# create a HPA for a deployment, min number of pod:1, max pods:10
kubectl autoscale deploy my-app --cpu=50% --min=1 --max=10
# horizontalpodautoscaler.autoscaling/my-app autoscaled

# confirm
kubectl get hpa
# NAME     REFERENCE           TARGETS       MINPODS   MAXPODS   REPLICAS   AGE
# my-app   Deployment/my-app   cpu: 0%/50%   1         10        1          21s

kubectl describe hpa my-app
# Name:                                                  my-app
# Namespace:                                             default
# Labels:                                                <none>
# Annotations:                                           <none>
# CreationTimestamp:                                     Fri, 31 Oct 2025 17:43:25 -0400
# Reference:                                             Deployment/my-app
# Metrics:                                               ( current / target )
#   resource cpu on pods  (as a percentage of request):  0% (0) / 50%
# Min replicas:                                          1
# Max replicas:                                          10
# Deployment pods:                                       1 current / 1 desired
# Conditions:
#   Type            Status  Reason               Message
#   ----            ------  ------               -------
#   AbleToScale     True    ScaleDownStabilized  recent recommendations were higher than current one, applying the highest recent recommendation
#   ScalingActive   True    ValidMetricFound     the HPA was able to successfully calculate a replica count from cpu resource utilization (percentage of request)
#   ScalingLimited  False   DesiredWithinRange   the desired count is within the acceptable range
# Events:           <none>

# update: 50% -> 30%
kubectl edit hpa my-app
# horizontalpodautoscaler.autoscaling/my-app edited

# confirm
kubectl get hpa
# NAME     REFERENCE           TARGETS       MINPODS   MAXPODS   REPLICAS   AGE
# my-app   Deployment/my-app   cpu: 0%/30%   1         10        1          3m6s

# delete
kubectl delete hpa my-app
# horizontalpodautoscaler.autoscaling "my-app" deleted from default namespace

---

Lab: Create HPA - Declarative method

# my-hpa.yaml
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: web-hpa
  namespace: default
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: app-deploy
  minReplicas: 2
  maxReplicas: 10
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 60

kubectl create -f my-hpa.yaml
# horizontalpodautoscaler.autoscaling/web-hpa created

kubectl get hpa
# NAME      REFERENCE               TARGETS              MINPODS   MAXPODS   REPLICAS   AGE
# web-hpa   Deployment/app-deploy   cpu: <unknown>/60%   2         10        0          32s

kubectl delete hpa web-hpa
# horizontalpodautoscaler.autoscaling "web-hpa" deleted from default namespace