Indexer import jobs API surface

• Status: Accepted
• Date: 2026-01-31
• Context:
  • Need REST coverage for indexer import jobs (create/run/status/results) to satisfy ERD indexer checklist.
  • Must preserve stored-procedure boundaries, stable errors, and testable handlers with E2E coverage.
• Decision:
  • Add import job request/response models and handler wiring for create/run/status/results endpoints.
  • Extend app facade mapping for import job error translation and results/status projection.
  • Update OpenAPI and Playwright API coverage for new endpoints.
  • Alternatives considered: defer API surface until full import pipeline; rejected to keep parity with checklist and procs.
• Consequences:
  • Positive outcomes: import job endpoints are now reachable, documented, and covered in E2E.
  • Risks or trade-offs: runtime worker support now finalizes jobs deterministically, but full live importer execution for all import kinds remains incremental.
• Follow-up:
  • Implement background import execution and UI flows for import job monitoring.
  • Extend CLI support once import pipeline is ready.

Task record

• Motivation: close the ERD indexer checklist gap for import job REST endpoints and E2E coverage.
• Design notes: handlers trim inputs, map stored-procedure error codes to stable API errors, and return typed models; no inline SQL added.
• Test coverage summary: added API E2E coverage for create/run/status/results; existing unit tests cover trimming and error mapping.
• Observability updates: no new spans or metrics required for handler-only changes.
• Risk & rollback plan: rollback by reverting endpoint wiring and OpenAPI updates; no migrations or data changes.
• Dependency rationale: no new dependencies added; reused existing models, handlers, and stored procedures.