On October 7, 2025, between 2:40pm EDT and 3:40pm EDT, all services on Duo 34 were unavailable. This was due to a misconfiguration in our edge load balancer which stopped all traffic to that deployment. Duo engineering promptly fixed the misconfiguration and have fixed the underlying cause, to ensure our systems are more resilient.
Duo’s application platform has two parts to manage our services. One-part handles creating the network setup, and the other controls how traffic reaches running services. A service disruption occurred because of a misalignment between these two key systems. Even though the network was set up correctly, changes were not reflected in the traffic routing settings. This caused traffic to stop reaching the services, resulting in the outage.
To restore service, a manual fix was put in place to connect these network servers with the right settings. As a long-term solution, we have updated the system settings to point to the right destinations. This will ensure that the misalignment will not happen again. To avoid this problem in the future, we're reviewing how we manage system configurations, improving our setup process to avoid local files, and adding checks to stop changes that could cause similar issues.