Gateway Timeouts on scaled out Azure App Services?

Urs Enzler 41

Our scenario: We have a simple client-server architecture where a PWA is using our Azure Web Service (Windows hosted) as the backend. The backend has scaling out enabled (min 2 instances). We have this setup for years and it worked well.

However, for about 1-2 weeks, the client has been receiving Gateway Timeouts from our backend. Although we cannot say this with absolute certainty, Gateway Timeouts seem to occur when a call to the backend is being made while an app service instance is shut down.

For us, it looks like the standard load balancer in front of our app service instances redirects the call to an already shut-down instance. This is a behavior we didn't see in the last years.

Can someone confirm this behavior and/or our hypothesis? And is there a way to avoid this?

Urs Enzler 41 Reputation points

2024-06-03T06:28:50.08+00:00

We found out that the problem is not related to scale-down operations. So, we have no clue at all...
Grmacjon-MSFT 17,136 Reputation points

2024-06-19T03:24:02.6233333+00:00

Hi @Urs Enzler are you still facing this issue?
Urs Enzler 41 Reputation points

2024-06-19T07:23:49.66+00:00

No, we found that these were not Gateway Timeout errors at all. Due to a wrong configuration, the Angular service worker was used instead of making the calls from the client directly to the backend. The Angular Service Worker maps any error to a Gateway Timeout Error.
So, after preventing normal calls to use the Service Worker, we got the correct errors and could solve them.

Share via

Gateway Timeouts on scaled out Azure App Services?