/k8s-health — cluster health sweep

Context: $ARGUMENTS (⚠️ never run bare kubectl — always pass an explicit --context.)

Invoke kubernetes-platform-ops and run the read-only sweep, then summarise findings by severity:

K="kubectl --context=<ctx> --request-timeout=30s"
$K get nodes -o wide
$K get pods -A -o wide | grep -iE 'CrashLoop|Error|Pending|ImagePull'   # CrashLoop keeps phase=Running!
$K get events -A --field-selector type=Warning | tail -20
$K -n argocd get applications                 # Unknown/OutOfSync/ComparisonError = GitOps not reconciling
$K get pv,pvc -A | grep -v Bound
$K get clusterissuers 2>/dev/null             # zero issuers = no TLS despite cert-manager
$K top nodes 2>&1                             # "Metrics API not available" = no metrics-server

Flag the known traps: no default StorageClass (Pending PVCs), MetalLB+Cilium pool overlap, cert-manager with zero ClusterIssuers, ArgoCD prune+selfHeal over Delete-reclaim PVs, shared root disk near full. Report evidence-first; propose fixes but do not mutate without a go-ahead.