Azure OpenAI 403 content violation

Jonathan McLean 40 Microsoft Employee

I am getting this error when trying to send anything to all Azure OpenAI models on one endpoint.

Forbidden: Your resource has been temporarily blocked because we detected behavior that may violate our content policy. For more details on Azure OpenAI service content policy, please visit https://aka.ms/aoaicodeofconduct

I do not believe I am violating the Code of Conduct. How do I find out what is causing this to be disabled?

Anshika Varshney 15,625 Reputation points Microsoft External Staff Moderator

2026-05-01T02:43:10.3166667+00:00

Hi J**onathan McLean,**Thank you for bringing this to our attention. We are currently checking the details and working on resolving the issue. We’ll update you as soon as we have more information.

Thankyou!
Karnam Venkata Rajeswari 5,340 Reputation points Microsoft External Staff Moderator

2026-05-01T03:48:15.77+00:00
Hello @Jonathan McLean

Welcome to Microsoft Q&A .Thank you for reaching out to us.

The error message “Your resource has been temporarily blocked because we detected behavior that may violate our content policy” indicates a temporary resource-level restriction applied by the Azure OpenAI service. This type of enforcement impacts all deployments under the same resource and is triggered based on aggregated safety signals rather than a single request.

This restriction is applied as a protective measure to ensure responsible usage. It may occur even when individual prompts appear compliant, as the system evaluates both:

Content signals (prompts and responses)

Usage patterns (traffic volume, repetition, automation behavior)

Content is classified across categories such as:

Hate or fairness

Violence

Sexual content

Self-harm

Requests that meet certain severity levels, or patterns that resemble high-risk usage, can contribute to a temporary block.

Since exact triggers are not directly exposed, a structured review can help identify contributing factors

Reviewing recent request activity

Analyze logs leading up to the issue

Identify repeated or similar prompts

Check for sudden spikes or batch/automated traffic

Checking prior safety signals

Please look for earlier responses that were:

Filtered or partially completed

Associated with content safety warnings

The repeated retries of similar inputs may strengthen enforcement signals

Analyzing traffic sources

Please confirm whether multiple applications share the same resource

Identify scripts or automation generating repetitive or high-volume requests

Review whether raw user input is being passed without preprocessing

Please check if the following resolution help -

As immediate actions , please see if

Temporarily pause traffic to allow system cooldown

Please avoid repeated retries during the blocked period

Monitor for automatic recovery, as some restrictions are lifted after a short interval

As preventive improvements ,

Consider introducing input validation and sanitization for user-generated content

Apply rate limiting to control request bursts

Separate development, testing, and production workloads across resources

Review automation logic for repetitive or unintended prompt patterns

The following references might be helpful , please check them out

Content filtering for Microsoft Foundry Models (classic) - Microsoft Foundry (classic) portal | Microsoft Learn

Configure content filters (classic) - Microsoft Foundry (classic) portal | Microsoft Learn

Transparency Note for Azure OpenAI in Microsoft Foundry Models - Microsoft Foundry | Microsoft Learn

Thank you
Karnam Venkata Rajeswari 5,340 Reputation points Microsoft External Staff Moderator

2026-05-06T08:06:00.3166667+00:00

Hello @Jonathan McLean ,

Checking in to see if the above response was helpful

Thank you
Karnam Venkata Rajeswari 5,340 Reputation points Microsoft External Staff Moderator

2026-05-06T19:14:01.3566667+00:00

Hello @Jonathan McLean ,

Checking in to see if you had any chance to review the above response.

Thank you
Jonathan McLean 40 Reputation points Microsoft Employee

2026-05-06T20:12:10.32+00:00

The endpoint is shared among multiple applications and users so tracking down individual prompt calls or logs is very difficult.

Is there a way to review logs, prompts, responses, and usage at the endpoint level instead requiring access to the end-user application or machine?
Anshika Varshney 15,625 Reputation points Microsoft External Staff Moderator

2026-05-08T05:00:29.5+00:00

Hi Jonathan,

Did you get any chance to review the response.

Thankyou!

Answer accepted by question author

Anshika Varshney 15,625 Microsoft External Staff Moderator

Hi Jonathan,

I understand how confusing this situation can be, especially when your prompts appear safe but the whole resource is still blocked.

This 403 error usually means your Azure OpenAI resource has been temporarily restricted due to aggregated safety signals, not necessarily a single request. The system evaluates both the content (prompts/responses) and usage patterns (like repeated requests, retries, or traffic spikes) before applying this type of restriction.

So even if individual prompts look compliant, earlier activity or overall request patterns may have contributed to the block.

What you can try next:

Check recent activity logs in Azure Monitor
- Look for any filtered responses or warnings before the block happened
Review request patterns
- Repeated or similar prompts
- High-frequency or automated calls
Pause traffic for some time
- Temporary blocks often auto-recover after a cooldown period
Avoid repeated retries during the blocked state

If the issue still continues, I recommend raising a support request from Azure portal and sharing the Apim-request-id from your error response. The support team can check backend signals and help identify the exact trigger.

Let us know if you can share more details (like request pattern or logs) in the private chat, and we can help you narrow it down further.

Thankyou!

Anshika Varshney 15,625 Reputation points Microsoft External Staff Moderator

2026-05-13T15:35:21.6666667+00:00

Hi Jonathan,

Thank you for sharing the update in the teams. I appreciate you taking the time to confirm the resolution! Since I’ve converted comment into an answer, could you please take a moment to mark it as Accepted? This helps others in the community with the same question find the solution more easily.

Thankyou!

Answer accepted by question author

Salamat Shah 750 MVP

Try this and check:

Go to Azure Portal → OpenAI resource → Monitoring → Logs/Metrics and check recent activity.
Look for high volume, rapid retries, or repeated prompts.
Reduce request rate and stop any automated spam-like calls.
Regenerate API key and test with a single request.
Wait a bit (temporary blocks often auto-clear).

Jonathan McLean 40 Reputation points Microsoft Employee

2026-05-06T21:24:45.35+00:00

That tells me the amount of recent activity but how can I see recent calls in more depth so I can determine which prompt calls are causing the content block or who is making them? By policy we do not use API keys so regenerating the API key will not help.

1 additional answer

Jonathan McLean 40 Reputation points Microsoft Employee

2026-05-13T14:12:51.15+00:00

I received these instructions from support offline. This was helpful information for me.

You can check all requests at resource level using Azure Monitor:

Enable Diagnostic logs → Log Analytics

Then query logs (KQL) to see request patterns, errors, volume, etc.

But limitation:

You cannot easily identify which app/user is sending requests if endpoint is shared.

For full visibility per app, best approach is to use APIM in front of OpenAI (to track caller, retries, etc.).
Was this answer helpful?

0 comments No comments
Sign in to comment

Use comments to ask for clarification, additional information, or improvements to the question.

Azure OpenAI 403 content violation

1 additional answer

Your answer