How do you avoid rabbit holes?

It's nice to go deep on a problem sometimes, but we can be tempted to go too deep. How do you manage your time when examining intricate problems?