This page used to contain a long “current state” assembled on 2026-06-19. It became misleading as GitOps adoption, bare-metal migration, Longhorn, workload storage migration, and Velero were completed. Volatile state now belongs in the generated cluster snapshot; durable design belongs in devata-system-map; reconstruction order belongs in reconstructing-devata; chronology remains in journey.

Source order

  1. Live read-only cluster inspection describes what is running now.
  2. The default branch of the lab repository describes intended state.
  3. devata-system-map describes durable architecture and ownership boundaries.
  4. The generated public snapshot exposes allowlisted runtime values with a freshness timestamp.
  5. journey and case studies explain how the current design was reached.

Verification on 2026-07-26

Read-only inspection observed:

  • three Ready Talos nodes: one control plane and two workers;
  • 17 Argo CD Applications;
  • Cilium as the dataplane without kube-proxy;
  • Longhorn installed with selected observability PVCs using its StorageClass;
  • Velero installed for bounded off-cluster backup;
  • no Helm releases outside Argo CD ownership.

One Application was OutOfSync during the check while remaining Healthy. That is an example of why a dated inventory is evidence, not a timeless claim. The diff must be investigated in the platform workflow rather than explained away in this page.

Refresh commands

kubectl config current-context
kubectl get nodes -o wide
kubectl -n argocd get applications.argoproj.io
kubectl get storageclass
kubectl get pvc -A
helm list -A

These commands are read-only. Complete any status claim with an end-to-end transaction relevant to the component being assessed.