Microsoft says the global Microsoft 365 and Exchange Online outage of November 12 was caused by a latent defect introduced in July through an update built for its networking gateway servers. In its preliminary post-incident report, dated November 15, the company said that at the worst point more than 80% of the affected gateway servers were unable to respond.

Three conditions combined to expose the bug. Teams automatically switched to the networking infrastructure used by Exchange Online about 15 hours earlier, which raised traffic on those servers, and roughly 14% of the gateway capacity was already unavailable because of a separate issue the previous day. Exchange Online could not fail away from the gateway the way Teams and other services could. Delayed mail was held rather than bounced.

Microsoft says the defect and its fix do not apply to broader Windows deployments, though some outlets describe it as a Windows bug. A final report is due on November 20.