Those bugs certainly contribute but a 5.4 second count @ 3.6 seconds is itself a problem - thats 9 seconds right there. It may be we're doing eager loading, as you say. setting a batch limit of 1 should eliminate that.
Those bugs certainly contribute but a 5.4 second count @ 3.6 seconds is itself a problem - thats 9 seconds right there. It may be we're doing eager loading, as you say. setting a batch limit of 1 should eliminate that.