Microsoft 365 outage drags into day two as server recovery stalls

Roughly 130,000 Downdetector reports flooded in about Outlook at the peak of a Microsoft 365 outage that has now stretched into its second day, after servers that engineers restored repeatedly became unresponsive again under live traffic.

The disruption began around midday Eastern Time on Thursday, November 12, and hit Outlook, Microsoft Teams, SharePoint Online, OneDrive, web-based Office apps, Microsoft 365 Copilot, Defender XDR and administrative tools. Businesses, government agencies, schools and individual users worldwide lost access to core workplace software.

What actually broke?

Microsoft blamed the incident on an unexpectedly large number of servers becoming unavailable after failing automated health checks. The remaining healthy servers absorbed the displaced workloads, and elevated pressure kept affected services from operating normally.

Notably, the company said user traffic was not the trigger. Infrastructure failures disrupted its ability to distribute workloads effectively — a capacity-and-orchestration problem, not a demand spike.

Engineers paused deployments, redistributed traffic and restarted affected servers, then tested an alternative health-check method to see if more capacity could safely come back. It didn't hold. In late-Thursday updates, Microsoft admitted the fixes were producing only patchy results.

"We are observing partial recovery across regions in which we've applied these adjustments; however, these actions aren't resulting in sustained recovery," the company said via its Microsoft 365 Status account.

That partial recovery meant users in different regions saw very different levels of disruption, even as problem reports gradually declined.

What was the real-world impact?

The outage reached well beyond email and chat:

• Australian organisations reported trouble processing invoices
• Some schools lost online learning sessions
• Government agencies and healthcare organisations reported service disruptions
• Businesses in several countries fell back on phone calls, Slack and physical meetings to coordinate work

Azure infrastructure appeared largely unaffected — which is itself the lesson. A failure in specific components of one ecosystem cascaded across dozens of dependent business applications.

What does this expose about single-vendor risk?

For organisations running on Microsoft 365, the incident laid bare an operational vulnerability: access to essential workplace functions can hinge on the stability of infrastructure controlled by a single provider. Business continuity plans that assume the cloud is always up — with no tested fallback for communications or access to critical documents — failed that assumption this week.

The financial cost wasn't immediately established, but prolonged loss of email, collaboration and document management typically delays transactions, slows decisions and cuts productivity, hardest in organisations with thin alternative systems.

Where does recovery stand?

As of early Friday morning UTC, Microsoft had given no firm timeline for full restoration. Engineers were still working to stabilise server capacity, restore affected infrastructure and improve traffic distribution. The company also hadn't disclosed a definitive root cause beyond the server failures and configuration factors still under investigation.

That leaves Microsoft with a challenge bigger than flipping services back on: proving the infrastructure behind millions of businesses can absorb failures without days of disruption. For product and ops teams, expect renewed pressure for multi-provider continuity planning.