To separate concerns, the journal could be leverage to schedule loading tasks:
```
|---------------------------------------------------------------+-----------------------------------------------------------------------------|
| pros | cons |
|---------------------------------------------------------------+-----------------------------------------------------------------------------|
| - separation of concerns (lister lists and creates origins, | - the task's policy (recurring, oneshot) is not part of the journal's event |
| no longer schedules | (that might be triggering a schema migration on the scheduler's part: new |
| - centralize the scheduling logic within that journal client | origin_type column in task_type table) |
| (that could simplify maintenance and be a step towards T1157) | |
|---------------------------------------------------------------+-----------------------------------------------------------------------------|
```