If you had trouble accessing Microsoft 365 apps between 12 and 14 November 2026, you were certainly not alone. A massive global outage knocked out core services, including Exchange Online, SharePoint, and Copilot, for millions of people worldwide.
Microsoft has now revealed the culprit behind the chaos: a sneaky bug that sat quietly inside its networking gateway servers for over three months.
According to Charles Lamanna, Microsoft’s executive vice-president of Copilot, agents, and platform, the issue boiled down to something called lock contention inside an internal process.
In plain English, that happens when too many tasks queue up for the exact same shared resource, causing the whole system to stall while everything waits its turn.
The issue stemmed from a July patch that triggered lock contention. A combination of extra traffic from a Teams failover, pre-existing capacity losses, and a separate failure in regional isolation caused over 80% of the gateway servers to crash within two hours during peak US traffic.
Because the latent bug could not be undone using standard rollbacks, engineers had to manually build and deploy a patch across huge blocks of servers, slowing down recovery, especially for Exchange Online, which was repeatedly swamped by backlogged traffic as servers came back online.

