US1 Hosting Issues

Incident overview

On Sunday evening, customer websites hosted on US1 (Central US) experienced service disruption following an unexpected virtual machine restart.

The restart coincided with an urgent Azure preventive maintenance event affecting underlying network infrastructure in the Central US region. During the reboot process, the affected server became stuck in a Windows restart/update loop, delaying service recovery.

Root cause

Azure identified a degradation in a network device connected to one or more virtual machines in the Central US region and initiated urgent preventive repair work to avoid wider unplanned failures.

As part of this event, virtual machines may be automatically rebooted or migrated by the Azure platform. In this case:

• The US1 virtual machine underwent an unexpected reboot
• During startup, Windows entered a restart/update loop
• The loop prevented normal service startup and temporarily blocked standard recovery actions

No data loss occurred. Operating system and data disks were retained throughout the event.

Incident response

• The issue was detected shortly after the unexpected restart
• Affected services were isolated to US1-hosted workloads
• Engineering investigated both platform health and recovery options
• The Windows restart loop resolved naturally once the underlying platform state stabilised
• Services began returning to normal and the platform was placed under monitoring

Once stability was confirmed, the incident was marked as resolved.

Mitigation steps

• Continued platform monitoring following recovery
• Verification of service availability and application health
• Review of Azure Service Health notifications related to Central US maintenance
• Confirmation that no data corruption or configuration loss occurred

Conclusion

This incident was caused by Azure-initiated urgent preventive maintenance combined with a Windows restart loop during VM recovery. While the initial restart was unplanned from an application perspective, it aligns with Azure’s advance notification of urgent infrastructure repair in the region.

All services are now operating normally. We will continue to monitor platform health and take proactive steps to reduce the impact of future infrastructure-level maintenance events.