Hourly monitoring reports improvements + new K008
This MR improves the hourly query reports and adds a new one.
Changes
-
Add
K008: Top queries by shared blocks hit+read- Ranks queries by (shared hit bytes + shared read bytes).
- Uses a new internal helper to do hourly topk by sum of two metrics (avoids the incorrect “union(top hits, top reads)” approach).
- Includes schema + unit tests + wiring into
generate_all_reports()and CLI--check-id.
-
Clarify
K003naming- Renames/clarifies
K003title to explicitly reflect total time = exec + plan (schema/title update).
- Renames/clarifies
-
Fix per-query-id extraction
- Improves extraction logic so per-query JSON generation correctly discovers queryids from reports (unit tested).
-
Misc hourly reporting adjustments
- Updates report names, query limits, and multi-cluster filename prefix behavior (docs/tests adjusted accordingly).
Files touched
reporter/postgres_reports.py-
reporter/schemas/K003.schema.json,K004.schema.json,K005.schema.json,K006.schema.json,K007.schema.json,K008.schema.json tests/reporter/test_extract_queryids_from_reports_unit.pytests/reporter/test_generators_hourly_unit.pytests/reporter/test_generators_unit.pytests/reporter/test_report_schemas_hourly.pyREADME.mdreporter/requirements-dev.txt