In app/jobs/users/follow_job.rb:5, I was conflicted between passing user_id, followable_id and followable_type to the active job (which seems to be what's most used in the codebase) and passing the objects user and followable directly. I opted for the latter, because it made things simpler/cleaner, but let me know if you think the former is more appropriate.

Hi @johncurcio, thanks a lot, unfortunately in this case the former is the way to go, for serialization and concurrency issues.

Between the queueing of the object and the moment the job is executed the resource could have disappeared from the database, which would mean that the job would fail indefinitely. We also don't want to serialize the object themselves

The pattern is: pass IDs, load the objects from the DB inside the job, exit the job if the objects don't exist and then perform the actual operation.

You can see a couple of good examples in the source code: https://github.com/thepracticaldev/dev.to/blob/master/app/jobs/articles/score_calc_job.rb and this is a good example of how to load an object by type and id https://github.com/thepracticaldev/dev.to/blob/master/app/jobs/notifications/notifiable_action_job.rb

I'm going to formally request changes, so that other reviewers will see the label