Azure Notification Hub is intermittently timing out. Is this a known issue?

Question

Today, I encountered an intermittent issue with the Azure Notification Hub when sending push notifications to both Apple and Android devices. It seems that, at times, the push notification service is taking between 30 and 60 seconds to complete the request. Using the test send feature in Visual Studio, I noticed that about 60-70% of the time, the push notification works instantly. The rest of the time, either the request times out (60 seconds is the timeout by default) or takes upwards of 30 seconds to complete.

Is this a known issue? It started happening over the weekend, according to my logs. Additionally, there was a problem with the Hub service back in December and again in May which was resulting in all requests to the Hub timing out. Those issues were reported here:
https://learn.microsoft.com/en-us/answers/questions/178623/send-a-test-push-notification-error-code-gatewayti.html?childToView=391041#comment-391041

Thanks,
dsav

Answer 1

Hi @dsav

We apologize for the delay in response and for the frustration this issue may have caused.

The test send feature uses the synchronous debug send API (the same API that’s invoked from the test send in portal). The regular send API which is documented is queued and processed asynchronously. The latency behavior between debug and normal send will not be the same due to this and some other implementation differences.

Please note, that send latency will vary based on load and we do not provide a guarantee on a delivery window. If you're experiencing extremely latency-sensitive, we recommend our direct send API.

We checked internally with the engineering team and this isn't a known issue. This issue requires deeper investigation. Please reach out to us at azcommunity@microsoft.com with the subject ATTN: Grace and in the body of the email include your Azure Subscription ID, region, namespace, and hub are you using

Thanks,
Grace

Answer 2

I think the issue has resolved itself, but I will update here if we run into the problem again.