Active – Azure OpenAI Service request failures in multiple regions

Impact Statement: Starting at 06:35 UTC on 10 March 2026, a subset of customers using Azure OpenAI Service in multiple regions may experience intermittent request failures or reduced availability. We detected this issue through automated service monitoring after observing an elevated rate of failed requests affecting a subset of Azure OpenAI model deployments across multiple regions. Our investigation determined the issue was caused by a recent change impacting specific model workloads. Mitigation actions have been completed in some regions, and service health is improving there. Recovery and rollback activities are continuing in the remaining affected regions while our engineering teams actively validate service health. Further updates will be provided in ~60 mins or as events warrant. 

Impact Statement: Starting at 23:20 UTC on 9 March 2026, a subset of customers using Azure OpenAI Service in multiple regions may experience intermittent request failures or reduced availability. We detected this issue through automated service monitoring after observing an elevated rate of failed requests affecting a subset of Azure OpenAI model deployments across multiple regions. Our investigation determined the issue was caused by a recent change impacting specific model workloads. Mitigation actions have been completed in some regions, and service health is improving in those regions. Recovery and rollback activities are continuing in the remaining affected regions while our engineering teams actively validate service health. Targeted notifications are being sent to affected customers through Azure Service Health. Customers can refer to Azure Service Health for subscription‑specific impact details and updates.Further updates will be provided within 2 hours, or as events warrant. 

Impact Statement: Starting at 23:20 UTC on 9 March 2026, a subset of customers using Azure OpenAI Service in multiple regions may experience intermittent request failures or reduced availability. We detected this issue through automated service monitoring after observing an elevated rate of failed requests affecting a subset of Azure OpenAI model deployments across multiple regions. Our investigation determined the issue was caused by a recent change impacting specific model workloads. Mitigation actions have been completed in some regions, and service health is improving in those regions. Recovery and rollback activities are continuing in the remaining affected regions while our engineering teams actively validate service health. Targeted notifications are being sent to affected customers through Azure Service Health. Customers can refer to Azure Service Health for subscription‑specific impact details and updates. Further updates will be provided within 2 hours, or as events warrant.