Is Templafy Down?

**Investigation** On **June 24, 2026**, 2:54 PM CET the engineering team identified an issue affecting the gateway infrastructure responsible for routing customer traffic. As a result, some requests were directed to service instances that were no longer available, causing intermittent request failures. The investigation determined that the gateway controller responsible for managing traffic routing did not consistently refresh its routing configuration following infrastructure changes, such as deployments or platform scaling events. This caused outdated routing information to remain active on some gateway instances, resulting in traffic being sent to endpoints that were no longer serving requests. The issue affected both **West Europe \(Production 0\)** and **West Europe \(Production 5\)** during separate deployment activities. **Mitigation** The engineering team immediately initiated mitigation by restarting the gateway instances responsible for traffic routing. Restarting these components forced the routing configuration to refresh, allowing traffic to be directed to the correct service instances. The mitigation successfully restored normal traffic routing for: * **West Europe \(Production 0\):** 2:54 PM CET – 3:08 PM CET * **West Europe \(Production 5\):** 3:22 PM CET – 3:42 PM CET Following the incident, the engineering team worked closely with the gateway vendor to investigate the observed behavior and evaluate long-term remediation options. **Resolution** The issue was resolved after the affected gateway instances were restarted and the routing configuration was successfully refreshed. As an additional precaution, the gateway component was subsequently downgraded to a previously proven stable version while the investigation with the vendor continues. **Post-Incident Actions** The engineering team completed a detailed review of the incident together with the gateway vendor to better understand the observed behavior. To reduce the likelihood of similar incidents, we have: * Reverted to a stable gateway controller version while the vendor continues its investigation. * Increased monitoring of gateway configuration synchronization during deployments and scaling events. * Continued collaboration with the vendor to validate future gateway releases before production rollout. **Impact and Scope** The incident affected a subset of customer requests hosted in the following production clusters: * **West Europe \(Production 0\):** June 24, 2026, from **2:54 PM CET** until **3:08 PM CET** * **West Europe \(Production 5\):** June 24, 2026, from **3:22 PM CET** until **3:42 PM CET** During these periods, some requests failed because traffic was temporarily routed to service instances that were no longer available. Once the gateway configuration was refreshed, normal service resumed. We sincerely apologize for the disruption caused by this incident. Providing a reliable and dependable service remains our highest priority, and we are continuing to work closely with our technology partners while strengthening our deployment and monitoring processes to help prevent similar issues in the future. --- [2026-06-24T16:42:14.958+02:00] (resolved) The incident has been resolved, and further information will be provided in a postmortem shortly. We apologize for the impact to affected customers. --- [2026-06-26T13:24:34.240+02:00] (monitoring) The incident has been successfully mitigated, and our team is actively monitoring the situation to ensure ongoing stability and performance. We are observing the systems to prevent any further disruptions. --- [2026-06-26T13:22:02.337+02:00] (identified) Customers using West Europe (Production 0) may have experienced errors or delays while loading the application. Further updates will be posted here soon.

# Investigation On June 8, 2026, at 10:01 AM CET, the engineering team became aware of delays and intermittent failures affecting AI-powered functionality within Document Agents, including outline preparation and document generation. The engineering team determined that the disruption originated from an incident affecting Microsoft Azure OpenAI Services across multiple regions. As Azure OpenAI is a dependency for AI Assistant functionality, increased latency and timeout rates in the external service impacted the processing of AI requests. While the platform continued to accept requests, slower response times caused a buildup of pending AI requests. As a result, some users experienced delays when generating outlines and documents, and in some cases workflows appeared stalled while requests were being processed. # Mitigation Following identification of the underlying cause, the engineering team focused on reducing delays and restoring normal processing capacity. The engineering team promptly implemented mitigation measures to reduce the impact of the Azure OpenAI service degradation and restore normal processing of AI requests. In parallel, Microsoft worked to restore normal Azure OpenAI service performance. These actions helped stabilize the service and reduce the impact on users while the external service disruption was being resolved. # Resolution The issue was resolved at 5:00 PM CET on June 8, 2026. Following Microsoft's remediation of the Azure OpenAI service disruption and the mitigation measures implemented by the engineering team, AI request processing returned to normal levels. After resolution, users were once again able to generate outlines and documents through Document Agents without abnormal delays or failures. The engineering team validated that service performance had returned to expected operating levels before closing the incident. # Post-Incident Actions To improve resilience and visibility into similar situations, the following actions have been identified: * Implement additional monitoring and alerting for AI service latency and timeout rates. * Add alerts for unusual increases in pending AI requests and processing delays. * Review scaling and recovery procedures to ensure faster response to external service disruptions. * Continue evaluating opportunities to improve the resilience of AI-powered functionality. # Impact and Scope This incident affected AI-powered functionality used by AI Assistant and Document Agents between 9:45 AM CET and 5:00 PM CET on June 8, 2026. A subset of users experienced delayed responses, failed AI requests, or workflows that appeared stuck while outlines or documents were being generated. The impact was caused by increased latency and timeout rates in Microsoft Azure OpenAI Services. We sincerely apologize for the disruption caused by this issue. Maintaining a reliable and seamless experience is our highest priority. We remain committed to delivering a dependable and high-quality service experience. --- [2026-06-08T16:45:46.722+02:00] (resolved) The incident has been resolved, and further information will be provided in a postmortem shortly. --- [2026-06-08T16:37:19.494+02:00] (monitoring) At 13:45 UTC on 08 June 2026, Microsoft confirmed that the platform issue affecting Azure OpenAI Service across multiple regions had been mitigated. --- [2026-06-08T16:38:35.284+02:00] (identified) Microsoft reported that, starting at 09:45 CET on 8 June 2026, it identified an incident affecting the Azure OpenAI Service in multiple regions. Further updates will be posted here soon.