Fanout amplifies downstream load

A push notification drives a 5x surge to your mobile home-screen API. Your home-screen service autoscales fine and stays healthy, but a shared downstream user-profile service, which it calls 8 times per home-screen request to hydrate widgets, saturates: profile p99 explodes, its DB hits connection limits, and home-screen requests start failing on the profile dependency. Profile is used by many teams and can't scale instantly. Dashboards show home-screen traffic up 5x but profile traffic up ~40x. How do you triage and mitigate without taking down profile for everyone?