Azure Open AI service unavailable east-us region

Question

I have Azure Open AI resource in east-us region and suddenly all our deployed models (gpt4o, gpt4o-mini, embedding models, o3-mini) stopped responding over API. Getting either 404 deployment not found or timeout error. The deployments exist, I checked. I went to AI foundry to Playground and tried the deployments from there, the same problem, either getting:
: Completions call failed. Please try again
or DeploymentNotFound: The API deployment for this resource does not exist. If you created the deployment within the last 5 minutes, please wait a moment and try again.

I have another resource in eu-sweden and this is working fine.

Answer 1

Hello @Marek Pavelka,

The issue impacting the Azure OpenAI service availability in the East US region has now been Mitigated. Services are operating normally, and you should no longer experience any disruptions.

If everything is working as expected on your end, we kindly request you to confirm and Accept the resolution. Please don’t hesitate to reach out if you continue to face any issues or have further questions.

Thank you!

Answer 2

Latest update: In Azure Service Health page there is:
Service: Azure OpenAI Service

Region: East US

Impact Statement: Starting at 02:30 UTC on 09 April 2025, you were identified as a customer using Azure OpenAI Service and Azure AI Services in East US, who may experience elevated latencies and availability issues across all AOAI models.  

Workaround: Customers can distribute their workloads across other regions, such as West Central US. 

Current Status: We determined that a recent deployment impacted backend instances upon which Azure OpenAI Service and Azure AI Services relies, which is leading to latency and availability problems. We are currently rolling back to a previous, known good build, which should bring our resources to a healthy state. The next update will be provided in 60 minutes or as events warrants.

Answer 3

Last update few minutes ago:

**
Current Status:** We determined that a recent deployment impacted backend instances upon which Azure OpenAI Service and Azure AI Services rely, which is leading to latency and availability problems. Our mitigation strategy consisted of performing a rollback to a previous, known build; however, it has not proven effective at alleviating impact. We are now focusing our efforts on identifying an appropriate mitigation path to resolve this issue. The next update will be provided in 2 hours or as events warrants.

Answer 4

Anyone interested can follow the status in their Azure Portal here: https://portal.azure.com/#view/Microsoft_Azure_Health/AzureHealthBrowseBlade/~/serviceIssues