Create a report runner

[Donnype]

Also see:

• https://github.com/minvws/nl-kat-coordination/pull/3352

Depends on:

• https://github.com/minvws/nl-kat-coordination/pull/3442

[underdarknl]

My thinking is as follows:

• Periodically fetch jobs from the scheduler. If there is a Job:
• Take the Job, which contains the Recipe ID, and organization.
• Connect to Rocky (via API), and initiate Rocky's Create report flow with Recipe and Organzation as arguments.
• Rocky creates report and stores as it currently does.

In the future, the Report job might be moved away from Rocky and moves to a lambda style runner that collects the info from Octopoes itself and reports to Bytes with the report-data.

[Donnype]

Agreed on the first two steps, but I do not see why we want an API call to rocky:

• Part of doing this asynchronously is avoiding timeouts, which will still be an issue if we make blocking API calls
• If we make the API call non-blocking, we are implementing some sort of background task/queueing system in rocky anyways, so why not pull from the scheduler API and start tasks in Rocky directly. By creating a worker in Rocky for instance to be consistent across services.

Also the question arises: which service does the API call? The boefjes so we can reuse the worker? Then we are creating quite a complicated detour for creating jobs. Or simply Rocky? Then why let a service call its own HTTP API.