ephracis
commented
8 years ago I suggest we create an ActiveJob that calls Project#refresh which pulls in more meta data and creates additional models such as commits, issues and users.
I suggest we use Sidekiq as a queue backend in production, but since that will require an additonal Redis database which will incur additional costs, we can live with the default in-memory queue during development to save money.
The in memory queue, while free, will mean we lose all unfinished background tasks when we deploy and the app has to restart.
When pulling in data from Github and Gitlab we should track more than just projects. We need to track commits and the developers making the commits.
When pulling in users we need to ensure we can separate "real" users from users we only fetched from the outside. A suggestion is to use the number of times the user has signed in, 0 means the user is a "ghost users" and thus have never signed in to Appatite.