What happened
Microsoft says the Microsoft 365 incident began at 1:26 p.m. Eastern on November 12, 2026. The service-health record was tracked as MO1550821 and covered Exchange Online, Microsoft Teams, SharePoint Online, Microsoft 365 Copilot and Microsoft Defender XDR. Customers in multiple regions reported connectivity, email and collaboration failures.
On November 13, 2026, Microsoft said the issue came from an earlier network-capacity upgrade that surfaced after a recent network change. Microsoft said it paused deployments, shifted traffic, restarted servers and restored capacity gradually. Microsoft extended the email timeout window to 48 hours while queues and backlogs recovered. Attribution posture: Microsoft says the outage was a technical issue and found no indication of malicious, cyber or AI-agent activity.
On November 14, 2026, Microsoft said service had been restored to the remaining affected customers and regions. Governments, healthcare services, courts, businesses, airlines and banks reported operational effects during the disruption. Microsoft said it would provide a full post-incident report; that report was not included in the cited sources available for this edition. The cited source did not publish the specific operational detail described as Full post-incident report.
Why this matters now
Microsoft says the event was not malicious, but it remains a security-leadership issue because communications, collaboration and Defender XDR availability can be dependencies during an unrelated incident. A cloud outage can therefore reduce both business operations and the tools used to coordinate response.
Service restoration is not proof that customer workflows recovered cleanly. Delayed messages, approval queues, customer notifications, court or healthcare processes and third-party integrations may require local reconciliation after the provider reports green status.
The outage also exposes concentration risk. Organisations that place email, collaboration, security telemetry, executive communications and workflow automation in one provider need an explicit failure design rather than an assumption that a global service will remain continuously available.
The decision for security leaders
Require business-continuity and security teams to test an out-of-band communications path that does not depend on Microsoft identity, email, Teams or SharePoint. Include executives, incident command, legal, communications and service-desk leadership.
Map workflows that accumulated delayed messages, approvals or notifications and assign reconciliation owners. Provider restoration should trigger local recovery validation rather than automatic closure of the business incident.
Use the post-incident report as an assurance gate. Compare the root cause, failure propagation and corrective actions with contractual resilience requirements, then document accepted concentration risk or fund specific alternatives.
Evidence of closure
- Out-of-band communications test reaches every required incident role.
- Dependency map identifies each critical workflow tied to Microsoft 365 availability.
- Reconciliation record closes delayed messages, approvals and notifications.
- Approved post-incident review records corrective actions or accepted assurance limitations.
The Security.io assessment
Microsoft’s preliminary explanation and explicit rejection of malicious activity reduce the need for cyber-incident attribution. They do not remove the need to preserve local impact records, because technical outages can still weaken security operations and regulated service delivery.
The incident demonstrates correlated failure across services that many organisations treat as separate controls. Email, collaboration, document access, AI integrations and security tooling can share infrastructure and identity dependencies even when procured as distinct capabilities.
Closure remains partly open until the full post-incident report explains the initiating change, why capacity was lost across regions, which safeguards failed and what corrective actions Microsoft has validated. Enterprises should not invent a deeper cause before that evidence is available.
Questions for the morning meeting
- Which critical workflows stopped because Microsoft 365 was unavailable?
- Can incident teams communicate and access runbooks outside the affected provider?
- Which security controls depend on Microsoft 365 service availability?
- What evidence or service commitments are required from the post-incident review?