# Expected Diagnosis: expensive-query-spike-during-incident-window

This case should start from TiDB expensive-query or slow-query evidence and then
converge on a read-hotspot path caused by an ineffective access path.

The expected answer should make these points explicit:

- For this alert type, TiDB-side expensive-query or slow-query evidence should
  be treated as the primary starting point.
- The final diagnosis should identify a large-scan or missing-access-path
  bottleneck behind the dominant expensive query.
- ANALYZE-related queries should be filtered out or clearly treated as secondary
  maintenance noise rather than the main culprit.
- Recommendations should prioritize index or access-path fixes, with statistics
  refresh only as follow-up.