ThreatBeaconXSubscribe
CriticalVulnerability Research · 5 min read

Microsoft 365 Global Outage Deepens as Restored Servers Become Unresponsive Under Traffic

Microsoft is facing a widespread Microsoft 365 service disruption after repeated recovery attempts failed to maintain stable infrastructure availability. The outage, which began on November 12, 2026, has continued into November 13, affecting users across multiple regions. The disruption initially involved Exchange Online mailbox connectivity failures but subsequently expanded to other Microsoft 365 services, including Outlook, Teams, SharePoint Online, Microsoft 365 Copilot, and Microsoft Defender XDR. Microsoft's investigation identified unexpectedly unavailable servers, failed automated health checks, and elevated infrastructure load as contributing factors. However, restarting affected servers has not consistently restored service. Some servers become unresponsive again after receiving user traffic. Microsoft is now pursuing alternative recovery measures, including traffic restrictions, traffic rerouting, server rebuilds using previous configurations, and infrastructure capacity improvements. The incidents are tracked under EX1550543 and MO1550821 in the Microsoft 365 Service Health Dashboard.

Written by ThreatBeaconX Research Team·Published Nov 13, 2026

The Microsoft 365 outage demonstrates the operational risks associated with large-scale cloud infrastructure dependencies.

Microsoft initially attempted to restore affected servers through infrastructure restarts and alternative health-check mechanisms. Although some servers successfully returned to service, the recovery was not sustained when normal traffic resumed.

The repeated server failures have complicated restoration efforts and resulted in intermittent availability across affected Microsoft 365 services.

Key developments include:

  • The disruption expanded beyond Exchange Online to multiple Microsoft 365 services and regions.

  • More than 120,000 Outlook problem reports were recorded by Downdetector at the reported peak.

  • Reports originated from 74 countries.

  • Microsoft confirmed that restored servers could become unresponsive again after receiving traffic.

  • Engineers are restricting selected traffic and rerouting connections to reduce infrastructure pressure.

  • A subset of affected servers is being rebuilt using a previous configuration.

  • Microsoft has not confirmed the underlying root cause or provided a firm timeline for complete restoration.

  • Microsoft has stated that it is not aware of malicious activity causing the outage.

Technical Description

The disruption originated within Microsoft's cloud infrastructure, where multiple servers unexpectedly became unavailable.

The initial investigation suggested elevated infrastructure load across several regions. Microsoft subsequently clarified that the increased load was not caused by a surge in user traffic. Instead, server availability issues shifted additional workloads onto the remaining operational infrastructure.

Microsoft also identified failures involving automated health checks, which determine whether servers are healthy and capable of processing requests.

An alternative health-check mechanism initially allowed a subset of affected servers to return to service. However, engineers later observed that some restored servers became unresponsive again after receiving user traffic.

This behavior prevented the initial recovery approach from delivering sustained service restoration.

Recovery Measures Underway

1. Traffic limiting

Microsoft is temporarily restricting selected traffic types from reaching affected infrastructure in certain locations. This is intended to reduce pressure on affected servers while recovery activities continue.

2. Server rebuilds

A subset of affected servers is being rebuilt using a previous configuration to determine whether reverting configuration changes can prevent the recurring unresponsive behavior.

3. Traffic rerouting

Engineers are adjusting traffic protocols, redistributing workloads, and using strategic traffic blocking to improve service availability across affected regions.

4. Infrastructure capacity improvements

Microsoft is applying capacity improvements where possible to support regions returning to service and accommodate traffic as restoration progresses.

5. Diagnostic investigation

Microsoft continues analyzing diagnostic data to identify the condition that causes restored servers to become unresponsive after receiving traffic.

Affected Microsoft 365 Services

The disruption has affected multiple Microsoft cloud services, although the impact varies by service and individual feature.

Service

Reported impact

Exchange Online

Mailbox connectivity failures

Outlook

Difficulty accessing inboxes and sending or receiving emails

Microsoft Teams

Disruptions to selected communication and collaboration features

SharePoint Online

Selected features affected by Exchange Online dependencies

Microsoft 365 Copilot

Disruptions to Copilot and access to work data

Microsoft Defender XDR

Issues accessing selected portal pages and solution tabs

Defender for Office 365

Potential Safe Links verification issues

Microsoft Planner

Selected plan operations affected

Microsoft To Do

Loading and task-management reliability issues

Microsoft Intune

Selected service capabilities affected

Microsoft Purview

Selected compliance, classification, and data-protection capabilities affected

Power BI

Intermittent failures involving certain report exports

Microsoft Graph

Requests involving Exchange Online data may time out

Not every feature within these services is necessarily affected. Administrators should refer to the individual incident communications for the latest feature-level impact.

Incident Timeline

November 12, 2026 – Initial disruption

Microsoft began investigating Exchange Online mailbox access failures affecting multiple connection methods.

November 12 – Infrastructure availability issues identified

Microsoft identified unexpectedly unavailable servers and increased load across the remaining infrastructure.

November 12 – Initial recovery attempts

Engineers began restarting affected infrastructure, redistributing traffic, and investigating automated health-check failures.

November 13 – Alternative health-check method tested

Microsoft validated an alternative health-check mechanism that allowed some affected servers to return to service.

November 13 – Recovery instability identified

Microsoft confirmed that some restored servers became unresponsive again after receiving user traffic. Initial recovery efforts did not provide the anticipated improvement.

November 13 – Alternative recovery measures introduced

Microsoft began limiting selected traffic, rerouting connections, and rebuilding a subset of affected servers using a previous configuration.

November 13 – Partial recovery observed

Microsoft reported partial improvements in regions where traffic-management measures were implemented. However, intermittent failures continued, and complete restoration remained unconfirmed.

Scale and Business Impact

Downdetector recorded more than 120,000 Outlook-related problem reports at the reported peak, with reports originating from 74 countries.

The disruption affected organizations relying on Microsoft 365 for daily business operations, communication, collaboration, and security administration.

Potential operational consequences include:

  • Loss of access to business email and delayed communications.

  • Disruptions to meetings, calendars, and collaboration workflows.

  • Reduced availability of cloud-based document and productivity services.

  • Delays in security monitoring and investigation workflows where Defender portal access is affected.

  • Interruptions to compliance, data-protection, and administrative operations.

  • Delayed processing of queued email after service recovery.

  • Disruption to third-party applications and integrations dependent on Microsoft 365 services.

The incident also affected some external integrations, including Microsoft 365 connectors used by third-party AI platforms.

IT Operations Recommendations

Immediate Actions

  1. Monitor incident EX1550543 and MO1550821 through the Microsoft 365 Admin Center or the alternative service-status portal.

  2. Communicate the outage and expected business impact to affected users.

  3. Avoid unnecessary password resets, Outlook profile rebuilds, or tenant-wide configuration changes when the issue is attributable to Microsoft's infrastructure.

  4. Monitor mail-flow delays and validate queued email processing after service recovery.

  5. Use approved alternative communication channels for critical business operations.

  6. Monitor Defender XDR availability and ensure alternative security-monitoring procedures are available if portal access is disrupted.

  7. Alert users about phishing emails, fake support calls, and fraudulent recovery instructions associated with the outage.

  8. Validate service availability from the organization's own tenant before declaring business operations fully restored.

Preventive Actions

  1. Establish business-continuity procedures for extended cloud service outages.

  2. Maintain alternative communication and collaboration arrangements for critical business functions.

  3. Review dependencies between Exchange Online and other Microsoft 365 services.

  4. Establish monitoring for email delivery, API availability, and cloud-service connectivity.

  5. Document operational procedures for security monitoring when cloud security portals are unavailable.

  6. Maintain incident communication templates for widespread third-party service disruptions.

  7. Conduct post-incident reviews to identify operational dependencies and recovery gaps.

Conclusion

The Microsoft 365 global outage highlights the challenges of restoring large-scale cloud infrastructure when recovered servers cannot consistently handle returning user traffic.

Microsoft's recovery strategy has shifted from basic server restarts toward traffic restrictions, alternative routing, and rebuilding selected infrastructure using previous configurations.

Although partial recovery has been reported, intermittent failures and the absence of a confirmed root cause mean organizations should continue monitoring service health and maintain business-continuity arrangements.

The incident also reinforces the importance of alternative security-monitoring procedures and phishing awareness during major cloud-service disruptions.