sw-kubernetes-architect_skill

Overview

This skill helps me act as a Kubernetes architect focused on generating safe, production-ready manifests one service at a time to prevent cluster-wide crashes. I combine GitOps patterns (ArgoCD/Flux), service mesh designs (Istio/Linkerd), and cloud provider best practices for EKS/AKS/GKE. The goal is repeatable, minimal-change rollouts and predictable scaling for enterprise workloads.

How this skill works

I inspect a single service’s requirements, current cluster topology, and GitOps pipeline settings, then produce manifest fragments and progressive rollout strategies. Manifests include resource requests/limits, readiness/liveness checks, probes, and safe deployment strategies (canary/blue-green/gradual). I also provide optional service-mesh and ingress configuration snippets and notes for ArgoCD/Flux integration.

When to use it

• Onboard a new microservice without risking cluster stability
• Migrate a service to a service mesh like Istio or Linkerd
• Bootstrap GitOps manifests for ArgoCD or Flux
• Harden deployments for EKS, AKS, or GKE production patterns
• Iteratively roll out configuration changes with minimal blast radius

Best practices

• Generate and review manifests per service, not entire application at once
• Set conservative resource requests and enforce limits; adjust from telemetry
• Use readiness probes and preStop hooks to ensure graceful shutdowns
• Prefer progressive rollouts (canary or gradual) for external-facing services
• Keep GitOps repositories declarative and single-source-of-truth for each service

Example use cases

• Create a Deployment, Service, and HorizontalPodAutoscaler for a new API service with conservative CPU/memory settings
• Add Istio VirtualService and DestinationRule for controlled traffic shifting during canary releases
• Convert a Helm-based service into GitOps-managed manifests for ArgoCD with health checks
• Generate AKS/EKS/GKE-specific settings such as nodeSelector, tolerations, and cloud provider load balancer annotations
• Produce a rollback plan and manifest diff to minimize downtime during config changes