Webhook herding

[cmelone]

Problem/Opportunity Statement

Gitlab webhooks may encounter issues finding our server from time to time, and we might not be able to collect the resource usage we want.

What would success / a fix look like?

Write up a cron job that grabs all the Gitlab jobs executed in the past N hours (thinking 24) and run them through our collection function.

I'm trying to think of a good way to preprocess the jobs we actually want to collect. If we took all the jobs, that's a lot of wasted DB queries on verifying that the job is already saved. We could just grab a list of all the job ids in the database, and eliminate those from consideration on being re-processed.