Overview
- A global outage shut down core Microsoft 365 services and left millions without email for more than 12 hours after network gateways failed on Nov. 12, 2026.
- Microsoft says the root cause was a software change made in July that created lock contention in gateway processes and only surfaced under high load on modern multi‑core servers.
- The company found that routine tests missed the bug because test hardware used older, lower‑core CPUs that did not reproduce the problem.
- Operational moves increased stress before the collapse: a Teams traffic switch back to a primary gateway and a prior incident that cut about 14% of gateway capacity shifted load patterns and helped trigger the cascade that affected roughly 80% of gateway servers.
- Microsoft published a preliminary incident summary on Nov. 16 and says a fuller technical report will follow later this week as engineers continue investigation and remediation, while third‑party monitors reported fresh, smaller service complaints on Nov. 17.