From 15:20 UTC to 15:26 UTC on February 21, 2019, the DUO1 deployment experienced a load balancer failure resulting in all authentication requests to that deployment failing. The issue was identified and resolved via automated remediation.
The DUO1 deployment’s load balancing tier is composed of several pairs of internally redundant load balancing hardware configurations. Each pair is composed of identically configured hardware in two separate locations in an active/standby setup.
At 15:20:21 UTC, one of DUO1’s active load balancers failed, preventing customers from connecting to DUO1. When this type of failure occurs, an automated process redirects customers to the standby load balancing hardware. The issue, which Duo detected immediately, was resolved automatically for all customers by 15:26:18 UTC. We will continue to explore additional infrastructure and software improvements to reduce failover time and customer impact moving forward.