Pin OpenWorkflow 0.11.0 in web, worker and shared jobs, with the matching lock update.
Use the native logger option and explicit queue-backend ownership. Reuse the worker's application DB and stop active work before closing its databases.
Keep operational logs free of raw engine payload/error arguments. Force-close health-only connections during shutdown so a stalled request cannot delay worker stop.
Remove custom OpenWorkflow history deletion. Preview-file/app-log cleanup remains application-owned; engine history stays until native retention is available.
This is the simplified upgrade, not a historical migration/recovery or retention framework.
Verification
Independent code/readability reviews and follow-up closures: clear.
All 23 checks passed; 2,417 tests passed, including explicit jobs tests/lint/typechecking.
Jobs/config/helpers/theme coverage: 100%. Worker line coverage: 98.34%, up from 97.89%.
Svelte: 0 errors, the same 27 existing warnings in seven untouched files.
Native fresh-queue startup/health/shutdown test includes active work and a stalled health connection. Existing HTTP/SSR and SDK-reader cases pass.
Config, archives, templates, OG compiler, publishing/preview/SSE contracts and protected samples remain unchanged.
Boundaries
The current startup enqueue and 24-hour timer remain in this PR. PR 2 will use native cron and durable jitter near 02:00 UTC, with focused seven-day preview-file/app-data cleanup.
No historical queue replay fixture, private engine-retention SQL, new service, rolling deployment framework or production Docker test. Existing database migrations remain library-owned. Engine history may accumulate until upstream retention is adopted. Production-image testing remains separately owner-owned.
## Summary
- Pin OpenWorkflow **0.11.0** in web, worker and shared jobs, with the matching lock update.
- Use the native logger option and explicit queue-backend ownership. Reuse the worker's application DB and stop active work before closing its databases.
- Keep operational logs free of raw engine payload/error arguments. Force-close health-only connections during shutdown so a stalled request cannot delay worker stop.
- Remove custom OpenWorkflow history deletion. Preview-file/app-log cleanup remains application-owned; engine history stays until native retention is available.
This is the simplified upgrade, not a historical migration/recovery or retention framework.
## Verification
- Independent code/readability reviews and follow-up closures: clear.
- All **23 checks** passed; **2,417 tests** passed, including explicit jobs tests/lint/typechecking.
- Jobs/config/helpers/theme coverage: **100%**. Worker line coverage: **98.34%**, up from 97.89%.
- Svelte: **0 errors**, the same 27 existing warnings in seven untouched files.
- Native fresh-queue startup/health/shutdown test includes active work and a stalled health connection. Existing HTTP/SSR and SDK-reader cases pass.
- Config, archives, templates, OG compiler, publishing/preview/SSE contracts and protected samples remain unchanged.
## Boundaries
The current startup enqueue and 24-hour timer remain in this PR. **PR 2** will use native cron and durable jitter near 02:00 UTC, with focused seven-day preview-file/app-data cleanup.
No historical queue replay fixture, private engine-retention SQL, new service, rolling deployment framework or production Docker test. Existing database migrations remain library-owned. Engine history may accumulate until upstream retention is adopted. Production-image testing remains separately owner-owned.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Summary
This is the simplified upgrade, not a historical migration/recovery or retention framework.
Verification
Boundaries
The current startup enqueue and 24-hour timer remain in this PR. PR 2 will use native cron and durable jitter near 02:00 UTC, with focused seven-day preview-file/app-data cleanup.
No historical queue replay fixture, private engine-retention SQL, new service, rolling deployment framework or production Docker test. Existing database migrations remain library-owned. Engine history may accumulate until upstream retention is adopted. Production-image testing remains separately owner-owned.