# Case: scale status

Use when `cluster_cr.status.phase` is `Scaling`, or user reports scale stuck /
instance-type scale unclear.

## Steps

1. Fetch `cluster_cr` (hard): phase, component replica / instance fields.
2. Fetch `scale_status_cm` via same `method-cluster-cr` call (hard if phase is
   Scaling; if absent when Scaling → note as signal / uncertain, do not invent).
3. Focus scale / instance-type scale phases (e.g. CreatingNodes, CordoningNodes,
   RestartingPods, DeletingNodes) and mismatch vs desired replicas.
4. APPBox must remain Active (hard Gate); if not → handoff `diag-appbox`.
5. Soft: DP Pods / Node readiness for the scaling component.
6. Provider-specific nodepool/nodegroup intents: see
   `diag-cluster/references/provider-*.md`.

## After recovery

Health checklist: phase back to `Running` + scale CM idle/absent as expected.