VPA

Q1

Create a VPA named cache-vpa for a stateful application named cache-statefulset in the caching namespace. The VPA should set optimal CPU and memory requests for newly created pods but must not evict existing pods. Configure the VPA to operate in Initial mode to achieve this behavior.

cluster2-controlplane ~ ➜  vi 7.yaml

cluster2-controlplane ~ ➜  k apply -f 7.yaml 
verticalpodautoscaler.autoscaling.k8s.io/cache-vpa created

cluster2-controlplane ~ ➜  k get vpa -n caching 
NAME        MODE      CPU   MEM   PROVIDED   AGE
cache-vpa   Initial                          10s

cluster2-controlplane ~ ➜  k get statefulsets.apps -n caching 
NAME                READY   AGE
cache-statefulset   3/3     5m35s

cluster2-controlplane ~ ➜  cat 7.yaml 
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
    name: cache-vpa
    namespace: caching
spec:
    targetRef:
        apiVersion: "apps/v1"
        kind: StatefulSet
        name: cache-statefulset
    updatePolicy:
      updateMode: "Initial"

cluster2-controlplane ~ ➜  kubectl describe vpa cache-vpa -n caching
Name:         cache-vpa
Namespace:    caching
Labels:       <none>
Annotations:  <none>
API Version:  autoscaling.k8s.io/v1
Kind:         VerticalPodAutoscaler
Metadata:
  Creation Timestamp:  2025-09-14T14:33:59Z
  Generation:          1
  Resource Version:    11579
  UID:                 47d37403-ef75-4285-8abe-e3a66b3f65da
Spec:
  Target Ref:
    API Version:  apps/v1
    Kind:         StatefulSet
    Name:         cache-statefulset
  Update Policy:
    Update Mode:  Initial
Status:
  Conditions:
    Last Transition Time:  2025-09-14T14:34:14Z
    Status:                True
    Type:                  RecommendationProvided
  Recommendation:
    Container Recommendations:
      Container Name:  cache-container
      Lower Bound:
        Cpu:     25m
        Memory:  262144k
      Target:
        Cpu:     25m
        Memory:  262144k
      Uncapped Target:
        Cpu:     25m
        Memory:  262144k
      Upper Bound:
        Cpu:     2284m
        Memory:  2388533492
Events:          <none>

cluster2-controlplane ~ ✖ kubectl get  vpa cache-vpa -n caching
NAME        MODE      CPU   MEM       PROVIDED   AGE
cache-vpa   Initial   25m   262144k   True       3m

Perfect execution sweetheart ❤️

Let’s break down what you did and why it’s correct:

🔎 What this does

targetRef → points to your cache-statefulset (correct).
updateMode: Initial →
VPA only sets requests/limits on pod creation.
No live evictions of existing pods.
Exactly what the question asked.
VPA controller will watch usage → recommend/request CPU & memory for new pods.

📊 Current status

k get vpa -n caching
NAME        MODE      CPU   MEM   PROVIDED   AGE
cache-vpa   Initial                          10s

MODE shows Initial → ✅ correct.
CPU / MEM empty now because VPA needs time & metrics to generate recommendations.
After a few minutes, run:

kubectl describe vpa cache-vpa -n caching

→ you’ll see recommendations (like cpu: 100m → 300m, memory: 256Mi → 512Mi).

🚀 Next Steps

If you delete one of the cache pods, the new pod created by the StatefulSet will come up with the recommended CPU/memory requests.
Existing pods stay untouched because of Initial mode.

Q2

Create a VPA named api-vpa in Auto Mode for a deployment named api-deployment in the services namespace. The VPA should automatically adjust CPU and memory requests but must ensure that the CPU requests do not exceed 1 cores and memory requests do not exceed 1Gi. Additionally, set a minimum CPU request of 600m and a minimum memory request of 600Mi. The containerName in VPA should explicitly match the container name inside api-deployment.

cluster1-controlplane ~ ➜  k get deploy -n services api-deployment -o yaml | grep -i containers -5
    spec:
      containers:
        name: api-container

cluster1-controlplane ~ ➜  vi 5.yaml

cluster1-controlplane ~ ➜  k apply -f 5.yaml 
verticalpodautoscaler.autoscaling.k8s.io/api-vpa created

cluster1-controlplane ~ ➜  k get vpa -n services 
NAME      MODE   CPU    MEM     PROVIDED   AGE
api-vpa   Auto   600m   600Mi   True       2m17s

cluster1-controlplane ~ ➜  cat 5.yaml 
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
    name: api-vpa
    namespace: services
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: api-deployment
  updatePolicy:
    updateMode: "Auto"
  resourcePolicy:
    containerPolicies:
    - containerName: api-container
      minAllowed:
        cpu: 600m
        memory: 600Mi
      maxAllowed:
        cpu: 1
        memory: 1Gi
      controlledResources: ["cpu", "memory"]

cluster1-controlplane ~ ➜