mark in-progress tickets as failed on daemon startup

[irmorteza]

Summary

• detect tickets still marked as in-progress, queued, or running
  when the daemon starts up
• mark those stuck tickets as failed with a clear message:
  "daemon restarted while ticket was running — rerun to continue"
• mark any associated running/queued jobs as failed as well
• skip tickets already in a terminal state (done, failed, cancelled)

Testing

• go test ./...

Notes

• recoverStuckTickets() is called once at daemon startup, before
  workers begin
• tickets can be re-run manually after recovery

Open Questions

• Auto-retry instead of fail? It might be better to automatically
  re-run stuck tickets instead of marking them as failed. However, this
  could cause an infinite loop if a ticket is fundamentally broken —
  the daemon would restart → re-run → crash → restart forever.
  A possible safeguard: add a retry_count field and only auto-retry
  if under a threshold (e.g. 3 attempts).