Description
PR #733 P2.10 remains in collectors/organization_collectors/api_usage.py:215-350. The primary overview can emit fresh status-code metrics while an optional multi-page operation enrichment leaves ORG_API_USAGE unmarked, causing a permanent stale alert and failure-retry cadence. Give the enrichment its own endpoint-group verdict or an explicit partial-success model so the scheduler and alerts describe each surface honestly.
Acceptance Criteria
- #1 Primary overview success refreshes the success signal used by its metrics and alert
- #2 Enrichment failure is separately observable and does not force the overview onto failure retry cadence
- #3 A regression test covers successful overview plus timed-out enrichment
- #4 Endpoint cost and cadence remain represented in the scheduler model
Definition of Done
- #1 make check (uv run ruff check . && uv run ruff format –check . && uv run mypy . && uv run pytest -v)
- #2 make docgen, when metrics, config, endpoints or collectors changed — CI fails the build on generated-docs drift
- #3 Grafana queries in grafana/dashboards/*.json and grafana/alerts/ updated, if a metric or label name changed
Implementation Plan
Wave 1 L3: split overview and enrichment endpoint-group outcomes with regression coverage; root integrates and finalizes.
Final Summary
Implemented and verified in 7327153. API-usage overview success is independent of optional enrichment failure; full gates passed.