Kubernetes Deployment

Kubernetes Deployment

This skill enables the agent to deploy and manage applications on Kubernetes clusters. The agent can generate deployment manifests, services, ingress rules, Helm charts, and autoscaling configurations. It handles the full lifecycle from initial deployment through scaling, rolling updates, and troubleshooting, following production best practices for resource management, security, and reliability.

Workflow

1. Configure Cluster Access: The agent verifies that kubectl is configured with the correct cluster context and namespace. It checks connectivity with kubectl cluster-info and confirms that the user has sufficient RBAC permissions to create and manage resources in the target namespace. If a kubeconfig is not present, the agent guides the user through authentication (e.g., aws eks update-kubeconfig, gcloud container clusters get-credentials).

2. Define Deployment Manifests: The agent creates Kubernetes deployment manifests specifying the container image, replica count, resource requests and limits, environment variables, liveness and readiness probes, and pod anti-affinity rules. Labels and annotations are applied consistently for service discovery, monitoring, and operations. The agent uses specific image tags (never latest) and sets imagePullPolicy appropriately.

3. Configure Services and Ingress: The agent creates Service resources to expose deployments within the cluster (ClusterIP) or externally (LoadBalancer, NodePort). For HTTP workloads, the agent configures Ingress resources with TLS termination using cert-manager, path-based routing, and rate limiting annotations. The agent selects the appropriate service type based on the deployment environment and traffic requirements.

4. Apply Manifests and Verify Rollout: The agent applies manifests using kubectl apply -f and monitors the rollout with kubectl rollout status. It verifies that all pods reach the Running state, health checks pass, and the service endpoints are registered. If a rollout stalls, the agent checks pod events with kubectl describe pod and logs with kubectl logs to diagnose the issue, and can execute kubectl rollout undo to revert to the previous version.

5. Configure Autoscaling: The agent sets up Horizontal Pod Autoscalers (HPA) to scale the replica count based on CPU utilization, memory usage, or custom metrics. It defines minimum and maximum replica counts, scale-up and scale-down behavior, and stabilization windows to prevent thrashing. For workloads with variable resource needs, the agent can also configure Vertical Pod Autoscalers (VPA).

6. Manage with Helm Charts: For complex applications with multiple environments, the agent packages Kubernetes manifests into Helm charts with templated values. Helm enables versioned releases, atomic upgrades with automatic rollback on failure, and environment-specific value overrides. The agent uses helm upgrade --install for idempotent deployments and helm diff to preview changes before applying.

Supported Technologies

• Orchestration: Kubernetes (EKS, GKE, AKS, self-managed), k3s, kind, minikube
• Package Management: Helm 3, Kustomize
• Autoscaling: HPA, VPA, KEDA, Cluster Autoscaler
• Networking: Nginx Ingress Controller, Traefik, Istio, Cilium
• Certificate Management: cert-manager, Let's Encrypt
• CI/CD Integration: ArgoCD, Flux, GitHub Actions, GitLab CI

Usage

Provide the agent with your application's container image, resource requirements, desired replica count, and target Kubernetes cluster details.

Example prompt:

Deploy my app to the production EKS cluster:
- Image: myregistry.io/myapp:v2.1.0
- 3 replicas with CPU/memory limits
- Liveness and readiness probes on /health
- Expose via Ingress at api.example.com with TLS
- HPA scaling between 3-10 replicas based on CPU

Examples

Example 1: Production Deployment with Service and Ingress

deployment.yaml:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: myapp
  namespace: production
  labels:
    app: myapp
    version: v2.1.0
spec:
  replicas: 3
  revisionHistoryLimit: 5
  strategy:
    type: RollingUpdate
    rollingUpdate:
      maxSurge: 1
      maxUnavailable: 0
  selector:
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
        version: v2.1.0
    spec:
      serviceAccountName: myapp
      terminationGracePeriodSeconds: 60
      affinity:
        podAntiAffinity:
          preferredDuringSchedulingIgnoredDuringExecution:
            - weight: 100
              podAffinityTerm:
                labelSelector:
                  matchExpressions:
                    - key: app
                      operator: In
                      values: [myapp]
                topologyKey: kubernetes.io/hostname
      containers:
        - name: myapp
          image: myregistry.io/myapp:v2.1.0
          ports:
            - containerPort: 3000
              name: http
          env:
            - name: NODE_ENV
              value: "production"
            - name: DATABASE_URL
              valueFrom:
                secretKeyRef:
                  name: myapp-secrets
                  key: database-url
          resources:
            requests:
              cpu: 250m
              memory: 256Mi
            limits:
              cpu: "1"
              memory: 512Mi
          livenessProbe:
            httpGet:
              path: /health
              port: http
            initialDelaySeconds: 30
            periodSeconds: 10
            timeoutSeconds: 5
            failureThreshold: 3
          readinessProbe:
            httpGet:
              path: /health
              port: http
            initialDelaySeconds: 5
            periodSeconds: 5
            timeoutSeconds: 3
            failureThreshold: 3
          lifecycle:
            preStop:
              exec:
                command: ["/bin/sh", "-c", "sleep 15"]

service.yaml:

apiVersion: v1
kind: Service
metadata:
  name: myapp
  namespace: production
spec:
  selector:
    app: myapp
  ports:
    - protocol: TCP
      port: 80
      targetPort: http
  type: ClusterIP

ingress.yaml:

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: myapp
  namespace: production
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-prod
    nginx.ingress.kubernetes.io/rate-limit: "100"
spec:
  ingressClassName: nginx
  tls:
    - hosts:
        - api.example.com
      secretName: myapp-tls
  rules:
    - host: api.example.com
      http:
        paths:
          - path: /
            pathType: Prefix
            backend:
              service:
                name: myapp
                port:
                  number: 80

Example 2: Horizontal Pod Autoscaler with Helm Deployment

Install or upgrade using Helm with custom values:

helm upgrade --install myapp ./charts/myapp \
  --namespace production \
  --set image.tag=v2.1.0 \
  --set replicaCount=3 \
  --set autoscaling.enabled=true \
  --values values-production.yaml \
  --wait --timeout 5m \
  --atomic

hpa.yaml — Horizontal Pod Autoscaler:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: myapp
  namespace: production
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: myapp
  minReplicas: 3
  maxReplicas: 10
  metrics:
    - type: Resource
      resource:
        name: cpu
        target:
          type: Utilization
          averageUtilization: 70
    - type: Resource
      resource:
        name: memory
        target:
          type: Utilization
          averageUtilization: 80
  behavior:
    scaleUp:
      stabilizationWindowSeconds: 60
      policies:
        - type: Percent
          value: 50
          periodSeconds: 60
    scaleDown:
      stabilizationWindowSeconds: 300
      policies:
        - type: Percent
          value: 25
          periodSeconds: 120

Deployment steps:

1. kubectl create namespace production (if it does not exist)
2. kubectl apply -f deployment.yaml -f service.yaml -f ingress.yaml
3. kubectl apply -f hpa.yaml
4. kubectl rollout status deployment/myapp -n production
5. kubectl get hpa myapp -n production to verify autoscaler targets