atlas-iac/infrastructure/core/node-maintenance.yaml

# infrastructure/core/node-maintenance.yaml
# Block new placement until offline, power and runtime-storage checks pass.
# Existing workloads and storage DaemonSets are not drained by these resources.
apiVersion: v1
kind: Node
metadata:
  name: titan-04
  labels:
    hardware: rpi5
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "Fresh undervoltage events; verify power delivery under load before accepting new workloads."
spec:
  unschedulable: true
---
apiVersion: v1
kind: Node
metadata:
  name: titan-05
  labels:
    hardware: rpi5
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "Offline; verify identity, power, storage and loaded operation before accepting new workloads."
spec:
  unschedulable: true
---
apiVersion: v1
kind: Node
metadata:
  name: titan-06
  labels:
    hardware: rpi5
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "Offline; verify identity, power, storage and loaded operation before accepting new workloads."
spec:
  unschedulable: true
---
apiVersion: v1
kind: Node
metadata:
  name: titan-08
  labels:
    hardware: rpi5
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "Runtime USB disk reaches sustained saturation during image builds; repair media and verify loaded I/O before uncordoning."
spec:
  unschedulable: true
---
apiVersion: v1
kind: Node
metadata:
  name: titan-14
  labels:
    hardware: rpi4
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    # Keep the runtime's other node fields; quarantine only owns placement.
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "Runtime USB flash saturation; replace or repair media and verify I/O before uncordoning."
spec:
  unschedulable: true
---
apiVersion: v1
kind: Node
metadata:
  name: titan-11
  labels:
    hardware: rpi5
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "Repeated fresh undervoltage events; verify power delivery under load before accepting new workloads."
spec:
  unschedulable: true
---
apiVersion: v1
kind: Node
metadata:
  name: titan-18
  labels:
    hardware: rpi4
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "Sustained runtime I/O stalls prevent Longhorn engine startup; reduce load and verify storage before uncordoning."
spec:
  unschedulable: true
---
apiVersion: v1
kind: Node
metadata:
  name: titan-12
  labels:
    hardware: rpi4
    node-role.kubernetes.io/worker: "true"
    longhorn-host: "true"
  annotations:
    kustomize.toolkit.fluxcd.io/ssa: Merge
    kustomize.toolkit.fluxcd.io/prune: disabled
    atlas.bstein.dev/maintenance-reason: "SD-backed runtime stalls caused correlated probe failures; verify loaded I/O before new placement."
spec:
  unschedulable: true