# Expected Diagnosis: map-selected-tikv-pods-to-worker-nodes-in-saas01

This ad-hoc case checks whether the agent can return the exact worker-node
mapping for a small set of live TiKV pods in the commonly used `saas01`
cluster.

The runtime hooks first resolve a recent stable snapshot, then dynamically pick
several active TiKV pods from that snapshot. The reference is built from live
`kube_pod_info` inventory for the same snapshot window.

The expected agent response should:

- map each selected `tikv_pod` to the correct `worker_node`
- preserve the same pod order as the prompt input list
- emit exactly one JSON code block with the required fields:
  `cluster_id`, `snapshot_time`, `evidence_sufficient`, `pod_mappings`
- avoid extra explanatory prose outside that JSON block

The key failure mode this case guards against is assigning a TiKV pod to the
wrong worker host.