Is Azure OpenAI Down?
Live Azure OpenAI status. Auto-refreshes every 2 minutes.
Azure OpenAI is Operational
Azure OpenAI is up and running normally. All systems are operational.
Official status page: azure.status.microsoft/en-us/status
Component Status
Recent Microsoft Azure incidents
What to do when Azure OpenAI is down
Fail over to another region or a comparable frontier API
Azure OpenAI is deployed per region, so your first move during an outage is to route traffic to a different region where you already have a deployment. If the disruption is broad or you have no spare region, switch to a comparable frontier API. The public OpenAI API runs the same model families on separate capacity, and Anthropic's Claude API is a strong drop-in for chat and reasoning workloads. Both are provisioned independently from Azure.
Route through a multi-provider gateway
If you do not want to hard-code a single backup, send requests through an inference gateway that can fail over across providers automatically. OpenRouter exposes hundreds of models behind one OpenAI-compatible endpoint, and gateways like Groq, Together, and Fireworks serve open-weight models on independent infrastructure so an Azure incident does not take your app down with it.
Get notified when status changes
Rather than refreshing this page during an incident, set an alert so you hear the moment Azure OpenAI flips back to operational (or goes down again). TensorFeed alerts watch the same feed and probes that drive this page, and the full status board shows every major AI provider side by side so you can pick a healthy fallback fast.
Frequently Asked Questions
Is Azure OpenAI down right now?
The live indicator at the top of this page is the answer. Green means the Azure status feed has no active items matching Azure OpenAI, Cognitive Services, or AI Foundry. Amber means Microsoft has published a degradation in at least one region, and red means a major service disruption. Because Azure OpenAI is deployed per region, a red or amber state often means a subset of regions is affected, not the whole service, so read the active items list for the regions and surfaces that are actually impacted.
How do I check Azure OpenAI status?
Three places. This page filters Microsoft's public Azure status feed (azure.status.microsoft) every two minutes for items mentioning Azure OpenAI so you skip the hundreds of unrelated Azure components. For incidents scoped to your own subscription and regions, open Azure Service Health in the Azure portal, which shows issues that the public page sometimes does not. You can also watch your deployment's own latency and 429 or 5xx error rates, which frequently move before any status item posts.
What should I do when Azure OpenAI is down?
Most Azure OpenAI outages are regional, so the fastest fix is to fail over to another region where you have a deployment, which is the recommended multi-region pattern. If the disruption is broad, route to a different provider such as the public OpenAI API or Anthropic's Claude API, since their capacity is provisioned separately from Azure. If you are seeing 429 throttling rather than an outage, that is a quota or rate-limit issue on your deployment, not a Microsoft incident, and you may need a Provisioned Throughput (PTU) deployment or a quota increase.
How often does Azure OpenAI go down, and how reliable is it?
Azure OpenAI is generally reliable and backed by an enterprise SLA, with most months passing without a customer-facing incident. When problems do happen they tend to be regional capacity or latency events rather than a global hard down, and many users in unaffected regions notice nothing. The most common day-to-day pain is not outages at all but 429 rate-limit responses during demand spikes, which is a quota constraint rather than a Microsoft service failure.
Which models and services does Azure OpenAI run, and how is it different from OpenAI direct?
Azure OpenAI hosts OpenAI's GPT family, the o-series reasoning models, text embeddings, and image and audio models like DALL-E and Whisper, with newer and third-party models increasingly offered through Azure AI Foundry. The key difference from OpenAI direct is that it is a separate product: separate endpoints (your-resource.openai.azure.com), separate Azure AD or key authentication, separate per-region capacity, and its own deployment types (Standard, Global Standard, Data Zone, and Provisioned). An OpenAI direct outage does not necessarily affect Azure OpenAI, and vice versa, because they do not share the same front-end capacity.
Where can I see Azure OpenAI incident history?
Microsoft publishes recent and resolved incidents at azure.status.microsoft, and the Azure portal's Service Health blade keeps a personalized history plus post-incident Root Cause Analysis (RCA) documents for issues that affected your subscription. The public status page only shows currently active items, so for past incidents the Service Health history view in the portal is the authoritative record.