# Expected Diagnosis: tikv-read-iops-ranking-in-premium-cluster-during-incident-window

This ad-hoc case evaluates a dynamic-window, cluster-scoped TiKV average
node-disk read IOPS top-3 ranking task for the latest completed Beijing
11:00-11:05 window.

The runtime hooks resolve the latest completed daily Beijing 11:00-11:05
window, then build a reference answer for that same window. The expected agent
response should:

- avoid treating a raw shared dataplane aggregate as directly cluster-scoped
- rank nodes by the average node-disk read IOPS over the full resolved
  5-minute window, not by a single endpoint snapshot
- emit exactly one JSON code block with the required fields:
  `cluster_id`, `window`, `mapping_strategy`, `evidence_sufficient`,
  `top_nodes_by_avg_read_iops`
- avoid extra explanatory prose outside that JSON block
- return the same cluster id and resolved window as the runtime reference
- use `tikv_node` as the ranked node identity in each array element
- rank the same top TiKV nodes by average read IOPS, in the same order, with
  numerically close `avg_read_iops` values

The judge intentionally avoids hard-coding tenant names or specific node names
in static YAML. Exact node identities and values come from the runtime-built
reference evidence.