# Expected Diagnosis: expensive-query-alert-correlation-during-incident-window

This case should conclude that a sustained expensive-query workload is the
primary problem, with TiKV latency and related alerts appearing as downstream
impact.

The expected answer should make these points explicit:

- The diagnosis should center on the expensive-query signal itself, not only on
  surrounding alert noise.
- A correct answer should connect the workload to TiKV-side latency or related
  downstream alerts instead of treating those alerts as independent root causes.
- Background conditions such as unrelated memory or TiFlash signals should be
  classified as secondary unless directly tied into the incident path.
- Recommendations should focus on isolating the dominant SQL digest or plan and
  optimizing that workload.