kubernetes（Kubernetes 集群运维）

When to use this skill

Use this skill whenever the user wants to:

• Write Deployment, Service, ConfigMap, Secret, or Ingress manifests
• Deploy, scale, or troubleshoot pods and clusters with kubectl
• Design resource limits, health probes, rolling updates, and operational workflows
• Set up local development clusters with minikube, kind, or k3d

How to use this skill

Workflow

1. Write manifests — define workloads and services in YAML
2. Apply to cluster — use kubectl apply -f to deploy
3. Verify status — check rollout, pod health, and service endpoints
4. Debug issues — inspect logs, describe resources, exec into pods

Quick Start Example

# deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: myapp
spec:
  replicas: 3
  selector:
    matchLabels:
      app: myapp
  template:
    metadata:
      labels:
        app: myapp
    spec:
      containers:
        - name: myapp
          image: myapp:1.0.0
          ports:
            - containerPort: 8080
          resources:
            requests:
              cpu: 100m
              memory: 128Mi
            limits:
              cpu: 500m
              memory: 256Mi
          livenessProbe:
            httpGet:
              path: /healthz
              port: 8080
            initialDelaySeconds: 10
          readinessProbe:
            httpGet:
              path: /ready
              port: 8080
---
apiVersion: v1
kind: Service
metadata:
  name: myapp
spec:
  selector:
    app: myapp
  ports:
    - port: 80
      targetPort: 8080
  type: ClusterIP

# Apply manifests
kubectl apply -f deployment.yaml

# Check rollout status
kubectl rollout status deployment/myapp

# View pod logs
kubectl logs -l app=myapp --tail=50

# Exec into a pod for debugging
kubectl exec -it deployment/myapp -- /bin/sh

Essential kubectl Commands

Best Practices

• Always set requests and limits for CPU and memory
• Configure livenessProbe and readinessProbe for every container
• Use Secrets for sensitive data and ConfigMaps for configuration
• Define rolling update strategy with maxSurge and maxUnavailable
• Collect logs and metrics centrally; use RBAC and NetworkPolicies in production

Troubleshooting

• CrashLoopBackOff: Run kubectl logs <pod> --previous to see crash output; check resource limits and probe configuration
• ImagePullBackOff: Verify image name/tag exists and imagePullSecrets are configured
• Pending pods: Run kubectl describe pod <name> — look for insufficient resources or unschedulable nodes
• Service not reachable: Verify selector labels match pod labels; check endpoints with kubectl get endpoints <svc>

Keywords

kubernetes, k8s, kubectl, deployment, pod, service, ingress, configmap, secret, container orchestration