We would love to know how other resolvers fared during the @facebook outage, and what their settings were. For example, did serve-stale appear to help anyone? In theory, this is the exact situation it was designed for...
(1/5) How @bind9 configuration defaults helped us :: Small deep-dive into recursive #DNS resolver at AS50242 during @Facebook outage incident.
Inbound rate of recursive DNS requests raised by 125%.