Implement periodic checking in on specific jobs.

- Managers should monitor the MQ for a new topic (`send_update` or something).
- Message should include a list of job ids to report on.
- For each job id, an `update` should be fired reporting the status or `lost` if Pulsar is no longer aware of the job.
- Galaxy should be configured to periodically send this request.

Galaxy with its "synchronicity", "actual database", "years or maturity", etc... is pretty good at not loosing track of jobs - so @natefoo and I decided it would be good if there was some way to periodically have Galaxy poll the jobs it believes to be open so if messages are lost or mishandled in transit the jobs do not stay open indefinitely on the Galaxy side.

Start by locating the manager MQ consumers and the existing handling for job status updates, then find how Galaxy configures polling of open jobs. Trace the current update and lost-job paths and determine where a periodic request containing job IDs can fit. Done means the request is handled for each listed job and Galaxy can periodically trigger it without leaving lost jobs open.

Written by the indexing model from the issue text.