The outage began last Thursday at 9 AM and lasted for nearly 24 hours, affecting Outlook users globally. The incident originated in Switzerland and spread to Mexico and other regions as failover systems attempted to reroute traffic. Within two hours, over 80 percent of Exchange Online servers became unresponsive.

Microsoft classified the incident as a sev0 event, the highest severity level. Hundreds of engineers participated in a conference call that reached the maximum capacity of 1,000 participants. The recovery was delayed because restored servers were immediately overwhelmed by traffic, creating a reboot loop.

The root cause was a latent bug in the networking gateway servers, present since July. The issue was triggered by a combination of factors: a Microsoft Teams failover that increased load and a 14 percent reduction in server capacity due to an unrelated prior issue. The bug caused lock contention in an internal Windows process.

Microsoft acknowledged that testing labs failed to identify the bug because they used hardware with fewer cores, whereas the issue primarily affected machines with more cores and newer hardware. The company is currently preparing a final post-incident review to be published later this week.