uncaged/workflow
10
0
Fork 0
Code Issues 6 Pull Requests Actions Packages Projects Releases Wiki Activity

fix: thread status detection — crashed threads stuck as 'running' #170

New Issue
Closed
opened 2026-05-09 12:19:45 +00:00 by xiaoju · 0 comments
xiaoju commented 2026-05-09 12:19:45 +00:00
Owner
Copy Link
Copy Source

Problem

When a workflow worker crashes (e.g. module resolution failure), the .running marker file is never cleaned up. The dashboard keeps showing the thread as "running" forever.

Example: Thread 06F0S8TQYHKR9Z829NNQAB3RZ8 failed but dashboard shows running.

Root Cause

packages/cli-workflow/src/serve/routes-thread.ts determines status purely by checking the .running file:

const status = r.source === "history" ? "completed" : isRunning ? "running" : "active";

If the worker process dies without cleanup, .running persists → stale "running" status.

Fix

  1. Status detection: Even if .running file exists, check the chain for a terminal node (__end__). If found → completed or failed, not running.
  2. Worker cleanup: Ensure the worker always removes .running on exit (add finally block or process signal handler).
  3. Distinguish completed vs failed: If the last node indicates an error, status should be failed, not completed.

Acceptance Criteria

  • Crashed/failed threads show correct status (failed)
  • Normally completed threads show completed
  • Only actively running threads show running
  • Worker cleans up .running file on any exit (success, error, signal)

小橘 🍊（NEKO Team）

## Problem When a workflow worker crashes (e.g. module resolution failure), the `.running` marker file is never cleaned up. The dashboard keeps showing the thread as "running" forever. **Example:** Thread `06F0S8TQYHKR9Z829NNQAB3RZ8` failed but dashboard shows running. ## Root Cause `packages/cli-workflow/src/serve/routes-thread.ts` determines status purely by checking the `.running` file: ```typescript const status = r.source === "history" ? "completed" : isRunning ? "running" : "active"; ``` If the worker process dies without cleanup, `.running` persists → stale "running" status. ## Fix 1. **Status detection**: Even if `.running` file exists, check the chain for a terminal node (`__end__`). If found → completed or failed, not running. 2. **Worker cleanup**: Ensure the worker always removes `.running` on exit (add finally block or process signal handler). 3. **Distinguish completed vs failed**: If the last node indicates an error, status should be failed, not completed. ## Acceptance Criteria - [ ] Crashed/failed threads show correct status (failed) - [ ] Normally completed threads show completed - [ ] Only actively running threads show running - [ ] Worker cleans up .running file on any exit (success, error, signal) --- 小橘 🍊（NEKO Team）
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
No items
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
aobing jiashuang (Jia Shuang) jiayi (Jiayi) luming scottwei tuanzi xiaoju xiaomo xiaonuo xingyue
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: uncaged/workflow#170
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?