in some cases the beaker-jobwatch does not provide the `job_ids=.*` pattern when it monitors beaker, that is because all jobs were already finished when running with `--allow-resume`.
```
beaker-jobwatch v8.1
Searching Beaker for jobs and clones to restore state...
Restored state: 0 running, 0 finished OK, 3 broken

2026-02-09 08:44:25
CKI tier0 - MR test level
Broken: 3
TJ#12301267 R:20636756 (x86_64): broken #3 (max): no systems match required HW
TJ#12301267 R:20636757 (x86_64): broken #3 (max): no systems match required HW
TJ#12301267 R:20636758 (x86_64): broken #3 (max): no systems match required HW
Running: 0/3
Completed: 3/3
no jobs in herd + herd_finished_ok, this should never happen
(unless you inspect finished broken jobs only)
[]
[]
'TJ#12301267 R:20636756', 'TJ#12301267 R:20636757', 'TJ#12301267 R:20636758'
duration: 0:00:00.001034
RecipeSets that finished successfully:
RecipeSets that did not finish: 17661123, 17661124, 17661125
finished unsuccessfully
```

Jira: CKI-6495