Question

I recently added User activity tracking functionality to my application which tracks user interaction with several models, and several different actions within each model.

class User

  has_many :activities, :order => 'created_at desc'     

  def log_activity(object, type)
    activities.create(:subject_type => object.class.name, :subject_id => object.id, :subject_action_type => type)
  end 

end

Now I call this method after a User creates/updates/deletes various objects and display this activity on the user profile. I imagine that after a while the Activities table will contain so many records it’s going to get to the point where it effects load times. I was wondering what a good implementation would be to manage this in a user-friendly way and keep my database running smoothly.

I was thinking of setting up some kind of scheduled task that cleans up old records after x amount of time. Perhaps you can give some advice as to how to accomplish this or recommend an even better solution?

I’m using PostgreSQL as my database of choice.

Answer 1

Your activities lazy loads so it won’t pull in all of the associated activities unless you ask for them. So when you say this:

activities.create(...)

All you’re doing to the database is a single INSERT: there is no select * from activities where ... to build activities, there’s just the INSERT from the create call.

So you don’t have anything to worry about from your log_activity method.

If you want to limit the size of your activities table then a simple:

Activity.connection.execute("delete from activities where created_at < now() - interval '10 days'")

every now and then should be fine to clear out the dead wood. The '10 days' is, of course, just for illustration.