BED-8747: Check-in to BHE during collection

[StranDutton]

OpenHound stops making any API calls to BHE while a collection job runs. BHE kills the job after the 10-minute client check-in timeout.

This PR adds a periodic check-in so BHE knows the collector is still alive throughout a collection.

Changes:

• Added a check-in thread to Service (src/openhound/scheduler/service.py) that calls GET /api/v2/jobs/current every 30s while a job is running in Service mode
• Check-in failures are caught and logged as warnings. They do not crash the thread or affect the running job.
• Check-in only: the response is ignored per the ticket (no status handling).
• Added unit tests

/resolves BED-8747