An instance routing egress traffic became unavailable, impacting a subset of SDK requests. A fix has been deployed and we're monitoring the system to ensure the system is fully operational.