Microsoft 365 outage lasted 23 hours 50 minutes — from November 12 at 16:20 UTC to November 13 at 16:10 UTC, and took down Exchange Online, Outlook, Teams, and OneDrive for businesses and governments worldwide. Microsoft confirmed it was not caused by cyberattack or AI agents. It was three small things, all real, all overlooked: a bug in the cloud gateway deployed in July.
A routine traffic shift that overloaded that gateway. And a 14 percent capacity shortfall across the server fleet that left no buffer when things went sideways. The result: more than 80 percent of affected gateway servers froze in lock contention. Regional isolation failed.
Recovery wasn’t a reboot, it was a patch, applied server by server, while engineers extended email delivery timeouts to stop delayed messages from bouncing. Services were fully restored by Monday morning. Microsoft’s final incident report is due November 20. It will address why regional fail-safes didn’t contain the cascade, and what changes are coming to prevent another outage where the ghost was already in the machine.




Be the first to comment
Join the argument. No password, just your email or a passkey.