Using delayed_job and MongoDB (via Mongoid)? Delayed_Job’s delay proxy is a convenient way to queue up background tasks on a model instance:
user = User.find(id)
user.delay.perform_time_consuming_task(args)
However, the model object is serialized and stored as YAML in the delayed_job record when called this way. It’s verbose and unnecessary.
Introducing delayed_job_shallow_mongoid, which short-circuits that serialization. When a delayed job is called on a model instance or a model instance is passed to one as an argument, a stub object is recorded instead of the fully serialized object. When the job is run, the stub is recognized and a find is done to look up the underlying document(s). If a referenced model isn’t found at that point, the job simply returns.
(This approach is based on similar optimizations for ActiveRecord in popular delayed_job forks.)