Kubernetes Platform Architect

Kubernetes has become the default substrate for running containerized workloads at scale, but running Kubernetes in production is significantly more complex than running a few demo clusters. The Kubernetes Platform Architect assistant helps platform engineering teams design Kubernetes environments that are production-ready from day one — covering cluster topology, multi-tenancy, networking, security, and the developer experience layer.

This assistant addresses managed Kubernetes services on all three major cloud providers: Amazon EKS, Azure AKS, and Google GKE. It helps you decide between single-cluster and multi-cluster architectures, design node pool strategies (including spot and preemptible node pools for cost optimization), configure cluster autoscaler and KEDA for workload-driven scaling, and establish network policies using Cilium, Calico, or cloud-native CNI plugins.

When you describe your workload profile, team structure, and operational requirements, the assistant designs a Kubernetes platform architecture tailored to your needs. This includes namespace and RBAC strategies for multi-tenant clusters, admission control with OPA Gatekeeper or Kyverno, GitOps delivery pipelines using Flux or ArgoCD, secrets management integration with Vault or cloud-native secret stores, and observability stacks using Prometheus, Grafana, and OpenTelemetry.

Expect detailed outputs including cluster topology recommendations with justification, resource quota and limit range strategies, ingress and service mesh architecture options (with Istio vs. Linkerd vs. Gateway API comparisons), container image supply chain security guidance, and platform upgrade and maintenance strategies. The assistant also helps design the internal developer platform layer — how application teams interact with the Kubernetes platform through abstractions like Backstage, Crossplane, or custom Helm chart libraries.

This assistant is ideal for platform engineers building internal developer platforms, cloud architects migrating workloads from virtual machines to containers, and SREs designing multi-tenant cluster environments. It is equally valuable for organizations adopting Kubernetes for the first time and those scaling from one cluster to dozens.