Investigation
The issue was detected at 10:22 AM CET on January 27, 2026, when Document Agents and AI Assistant services began experiencing degraded performance across the West Europe production environments (Production 0, Production 1, and Production 5). The engineering team began an immediate investigation after observing a pattern of failed requests and error responses affecting the document agent functionality.
The root cause was a service interruption in the OpenAI services running on our cloud provider, Microsoft Azure.
Mitigation
The team worked to contain the impact by closely monitoring service health and coordinating with Microsoft.
Resolution
By 6:37 PM CET on January 27, 2026, Microsoft had restored stability in the affected region. As the platform stabilized, document agents in West Europe (Production 0), West Europe (Production 1), and West Europe (Production 5) returned to normal operation and full functionality was restored.
Impact and Scope
This incident affected Document Agent and AI Assistant functionality for tenants in West Europe (Production 0), West Europe (Production 1), and West Europe (Production 5) between 10:22 PM CET and 6:37 PM CET on January 27, 2026. Users experienced intermittent availability issues during this period. The issue was isolated to these clusters and did not impact other regions.