Kubernetes Cluster Administration

Purpose

Design, deploy, secure, and operate production-grade Kubernetes clusters covering architecture decisions, networking, storage, security, scheduling, upgrades, and disaster recovery.

Agent Protocol

Trigger

Exact user phrases: "Kubernetes", "K8s", "kubeadm", "cluster", "kubectl", "etcd", "kubelet", "kube-proxy", "CNI", "Container Network Interface", "CSI", "Container Storage Interface", "RBAC", "PodSecurity", "PodSecurityPolicy", "PSA", "taint", "toleration", "node affinity", "topology spread", "cluster upgrade", "kubeadm upgrade", "cluster backup", "etcd backup", "Velero", "node pool", "karpenter", "cluster autoscaler", "control plane", "kube-apiserver", "kube-scheduler", "kube-controller-manager", "coredns", "network policy".

Input Context

• Cluster deployment method (kubeadm, kops, managed service, Talos)
• Kubernetes version (current + target for upgrades)
• Network plugin (Calico, Cilium, Flannel, Weave)
• Number of nodes + instance types
• Workload types (stateless, stateful, batch, GPU)
• Compliance requirements (PCI, HIPAA, SOC2, FedRAMP)
• Existing tooling (Helm, ArgoCD, Prometheus, cert-manager)