Flex Consumption Function App App Configuration references failed with OtherReasons for 2 days, then self-recovered

Question

We have an Azure Function App running on the Flex Consumption plan in UK South. This function app has been deployed for ~2 years now, and hasn't been edited in around the same amount of time. On 15 June 2026 at approximately 07:32 UTC, the app suddenly became unusable as all application settings using Azure App Configuration references failed to resolve.

In the Function App configuration view, the affected settings showed status:

OtherReasons

The underlying issue appeared to be that the App Configuration references could not be fetched/resolved. Some of these were Key-Value types, and others were Key-Vault references from the App Configuration resource.

This continued throughout 15 June and 16 June 2026. Restarting the Function App did not resolve the issue.

We also attempted to remove the system-assigned identity that is used, and re-add it, making sure to update the RBAC roles required, but this again resulted in the same error.

I was able to create a new Function App and give it the exact same environment variable references and that function app referenced them without issue, it appeared as if just this one single function app was affected.

As a mitigation on 16 June, we replaced the App Configuration references with direct environment variable values, which allowed the app to run again. We intentionally left one non-critical test App Configuration reference in place, pointing to a simple/random App Configuration value. That test reference continued to fail for the rest of the day.

On the morning of 17 June 2026, the test App Configuration reference started resolving successfully again without any change from our end. We then reverted the other settings back to App Configuration references, and the app is now working again.

The Function App is now working again with App Configuration references restored. However, we do not understand why reference resolution failed for this specific Function App for approximately two days and then recovered by itself.

Can we investigate this Function App and confirm:

• Why App Configuration references returned OtherReasons and failed to resolve between 15 June 2026 07:32 UTC and the morning of 17 June 2026?
• Whether there was any platform-side issue, backend incident, stale cache, identity resolution issue, or App Service/Flex Consumption configuration-reference issue affecting this specific Function App.
• Whether there are known issues or limitations with App Configuration references on Flex Consumption Function Apps.
• Whether any internal diagnostics are available for the failed reference-resolution attempts?

What mitigation Microsoft recommends to prevent a Function App becoming unable to start if App Configuration references temporarily fail to resolve.

Let me know if I could provide you any extra information with this! I can provide resource IDs on request.

Thanks

Answer 1

Hi DC,

Thank you for reaching out to Q&A portal. We have received update from backend regarding the root cause of the issue and resolution. Please find below and let me know if any further questions. 

 

The Microsoft Azure Team has investigated the issue you reported regarding Key Vault reference resolution failures for your Function App in the UK South region starting on 6/16/2026. This issue was found to be related to an internal platform infrastructure inconsistency where a portion of the components responsible for resolving Key Vault references on behalf of customer applications became unavailable. The platform's routing mechanism did not detect this unavailability promptly and continued to direct resolution requests to the affected components, causing Key Vault reference resolution attempts to fail intermittently.

  Upon investigation, engineers identified that the affected infrastructure components were no longer in a healthy state but had not been removed from the routing path. A targeted recovery action was performed to restore the Key Vault resolution service to full capacity in UK South, after which all Key Vault references resumed resolving successfully.   We are continuously taking steps to improve the Azure Functions service and our processes to ensure such incidents do not occur in the future, and in this case it includes (but is not limited to):

• Improving health detection to automatically identify and remove unhealthy resolution components from the routing path.
• Deploying additional monitoring to detect when resolution capacity drops below expected thresholds, enabling faster detection and response.
• Implementing automated self-healing capabilities so the platform can recover from this class of issue without manual intervention.

We sincerely apologize for any inconvenience.