kanban-dispatcher-stalled

Kanban Dispatcher stalling often stems from missing launchd plist files or script path mismatches. Diagnosing these issues involves verifying launchd trigger presence and validating that script files reside in the correct operational path to restore normal workflow execution.

To check for stale 'last_status=ok' markers in jobs.json, validate the marker's current state against active worker processes. This diagnostic step identifies inactive tasks incorrectly marked as successful, preventing dead worker reclaim failures and ensuring accurate recurring job scheduling.

Dead worker reclaim failures occur when inactive worker processes are not properly detected and reclaimed. Diagnosing this requires inspecting jobs.json and cron outputs to identify orphaned workers, allowing the dispatcher to clear stale status markers and resume sequential ticking.

Yes, diagnosing recurring job next_run_at issues requires python3 and sqlite3 to inspect the database and validate scheduling data. Access to launchctl and jobs.json outputs is also necessary to ensure recurring jobs have valid next run values and prevent sequential tick blocking.

An LLM cron job blocks sequential ticks by holding the dispatcher lock longer than expected. Diagnosing this involves analyzing cron outputs and jobs.json to detect the blocking process, resolve the deadlock, and restore sequential task progression.