More Than 80% of Gateways Failed and Exchange Could Not Step Aside

Microsoft's preliminary report says that at the worst point of the 12-13 November outage, more than 80% of the networking gateway servers behind Exchange Online were unresponsive. Failed servers pushed their traffic onto healthy ones, which failed too, and within two hours the defect showed up across the fleet. Teams, OneDrive and SharePoint could fail out of the faulty infrastructure. Exchange Online could not, so mail access and delivery suffered for close to 24 hours, according to The Register and BleepingComputer.

A Dormant July Defect Met Two Ordinary Events

The defect came in with a July update to the gateway servers and sat quiet for months. It caused lock contention in an internal Windows process that used up CPU and memory, and it got worse on machines with more cores. Two events set it off: Teams failed over onto the gateway infrastructure that serves Exchange Online, and almost 14% of gateway capacity was already offline for an unrelated reason. Microsoft says its load tests missed the bug because they ran on weaker hardware than production. A rollback was impossible because the change was months old, so engineers built a patch instead.

Plan a Mail Fallback Before the Final Report on 20 November

Microsoft promises more capacity, better monitoring and safeguards so that redirected traffic cannot cascade again, and it paused all planned deployments. Do not wait for that. Name the fallback path your team uses when Exchange Online is down for a day, test it once, and check what your contract says about credits for an outage this long. In the EU, NIS2 asks essential and important entities to keep business continuity measures in place, and a single mail supplier that cannot step aside is exactly the dependency an auditor will ask about.