Microsoft on Thursday updated Azure status history page to report that Azure storage availability and provisioning services have been restored following an outage in 26 out of the 28 regions of the public cloud.
First impact which occurred because of an incident in the East US region began at 21:50 UTC on Wednesday and continued till 06:00 on Thursday, affected Storage, services and customers relying on Storage and made provisioning of new resources or accessing existing resources difficult for them. Services that were affected included Azure Cache, Azure Logic Apps, Azure Stream Analytics, Azure Service Bus, SQL Database, Application Insights, API Management and more. Preliminary cause has been suggested as a Storage cluster that lost its power and was rendered unavailable for usage.
Second impact began at 22:42 UTC on Wednesday and continued till 00:00 UTC on Thursday and is said to have occurred because of an underlying storage incident, some Azure services utilizing Storage experienced issues related with service management. Because of this, only service management operations were affected, and not Storage resources. It may have caused VM or cloud services customers to experience failure while provisioning resources. Impacted customers of Storage could not have provisioned new resources or do service management operations on existing ones. Customers using services of Azure Search, Azure monitor, Azure Site Recovery, API Management, Azure Batch, EventHub were impacted with this incident. Azure Engineers have recognized a software error as probable root cause and a patch has been applied for mitigating this issue. Also, they observed approximately 50% success rate during the window of impact and so most customers would have been successful in retries.
For both the issues, Microsoft will publish a detailed root cause analysis within approximately 72 hours.