[Feature] Add queue/worker checks covering task publishing, worker startup, retry behavior, stale worker detection, and graceful shutdown #24

Closed
opened 2026-07-06 11:30:01 +02:00 by zemion · 2 comments
Owner

Imported from the consolidated GovOPlaN product backlog.

  • Source: /mnt/DATA/Nextcloud/ADD ideas UG/Products/govoplan/backlog.md
  • Line: 156
  • Section: Milestones > Milestone 8 - Production-Realistic Dev And Deployment Foundation
  • Source status: OPEN.

Imported backlog item:

- [ ] Add queue/worker checks covering task publishing, worker startup, retry behavior, stale worker detection, and graceful shutdown.
<!-- codex-backlog-fingerprint:d9e21e4fe411473acf968c61 --> Imported from the consolidated GovOPlaN product backlog. - Source: `/mnt/DATA/Nextcloud/ADD ideas UG/Products/govoplan/backlog.md` - Line: `156` - Section: `Milestones > Milestone 8 - Production-Realistic Dev And Deployment Foundation` - Source status: `OPEN.` Imported backlog item: ```markdown - [ ] Add queue/worker checks covering task publishing, worker startup, retry behavior, stale worker detection, and graceful shutdown. ```
zemion added this to the Milestone 8 - Production-Realistic Dev And Deployment Foundation milestone 2026-07-06 13:03:38 +02:00
Author
Owner

Substantial slice published, but this issue remains open for deployment proof.

Implemented:

  • Core asserts that every routed task queue is present in the configured worker queue list.
  • All deployment/dev defaults now include Mail, Calendar, Dataflow, Workflow, Postbox, Events, IDM, and default queues.
  • Late acknowledgement, worker-loss rejection, and prefetch-one delivery settings are regression tested.
  • Ops now detects stale/no heartbeat, missing queue consumers, active task count, and Redis queue depth.

Commits: GovOPlaN/govoplan-core@b65b488, GovOPlaN/govoplan-ops@05ce4dc, GovOPlaN/govoplan@3c658fa.

Remaining before closure: a real broker publish/consume probe, target retry/redelivery drill, and supervised worker drain/SIGTERM/graceful-shutdown acceptance test. Those require a running Redis/worker deployment rather than another static unit test.

Substantial slice published, but this issue remains open for deployment proof. Implemented: - Core asserts that every routed task queue is present in the configured worker queue list. - All deployment/dev defaults now include Mail, Calendar, Dataflow, Workflow, Postbox, Events, IDM, and default queues. - Late acknowledgement, worker-loss rejection, and prefetch-one delivery settings are regression tested. - Ops now detects stale/no heartbeat, missing queue consumers, active task count, and Redis queue depth. Commits: GovOPlaN/govoplan-core@b65b488, GovOPlaN/govoplan-ops@05ce4dc, GovOPlaN/govoplan@3c658fa. Remaining before closure: a real broker publish/consume probe, target retry/redelivery drill, and supervised worker drain/SIGTERM/graceful-shutdown acceptance test. Those require a running Redis/worker deployment rather than another static unit test.
zemion added
status
in-progress
codex/ready
and removed
status
triage
codex/needs-human
labels 2026-07-31 22:54:52 +02:00
Author
Owner

Implemented and verified.

Evidence:

  • Meta worker lifecycle drill: tools/checks/worker-runtime-drill.py.
  • Gitea Module Matrix run 300 reached and passed Prove worker delivery and shutdown guarantees against the Redis service and wrote audit-reports/worker-runtime.json.
  • The drill took a real worker through registration, delivery acknowledgement, graceful drain/shutdown, and redelivery-safe recovery boundaries.
  • Run: https://git.add-ideas.de/GovOPlaN/govoplan/actions/runs/300

The workflow failed only later in the independent WebUI module-permutation step because the installed Access package did not expose src/module.ts; that does not invalidate the completed worker evidence.

Implemented and verified. Evidence: - Meta worker lifecycle drill: `tools/checks/worker-runtime-drill.py`. - Gitea Module Matrix run 300 reached and passed **Prove worker delivery and shutdown guarantees** against the Redis service and wrote `audit-reports/worker-runtime.json`. - The drill took a real worker through registration, delivery acknowledgement, graceful drain/shutdown, and redelivery-safe recovery boundaries. - Run: https://git.add-ideas.de/GovOPlaN/govoplan/actions/runs/300 The workflow failed only later in the independent WebUI module-permutation step because the installed Access package did not expose `src/module.ts`; that does not invalidate the completed worker evidence.
zemion removed the
status
in-progress
codex/ready
labels 2026-08-03 02:32:17 +02:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: GovOPlaN/govoplan-core#24