Log non-200 OK responses separately and alert accordingly

[stevejalim]

At the moment, we call raise_for_status() after checking a URL to ensure it's a 200. This also is a crude way of catching 404s and 500s in the site -- if the status-check fails, the overall batch/job fails. So, it gets communicated because of the error, but it's a bit sledgehammer vs nut.

Instead, we should gather non-200 OK responses, log them to an output file and then report/fail/alert on those in the final reporting step.