How to fix web app deployment not loading - 504.0 GatewayTimeout

Dio 0

Hello,

My web app does not load and instead times out after a few minutes to the 504.0 Gateway Timeout error.

I am using the Azure Database for PostgreSQL flexible server and Azure Plan as the subscription and Node as the Runtime stack. Perhaps it is due to the scaling and I am using the free plan?

Here is my website! Please let me know if you need anything else

https://plan-it-matt-cha.azurewebsites.net/

1 answer

Dan Rios 1,265 Reputation points MVP

2024-03-18T21:07:24.3133333+00:00

Hi

The website seems to load for me at the moment.

However I would follow these steps to try and see what errors there are for you to look into further:

Theres a great native troubleshooting tool you can leverage to help identify your issues.

Go to your Azure app and on the left navigate down to Resource Health > Diagnose and troubleshoot button in this view

From here you should be able to get an overview of the errors with suggested fixes for these.

You can read more here:

https://learn.microsoft.com/en-us/azure/app-service/overview-diagnostics

If you have any further questions or need any further help drop a reply and I’ll be happy to try and assist further. If this was helpful in solving your issue then please mark it as an accepted answer so others can benefit from the resolution.
Please sign in to rate this answer.
Dio 0 Reputation points

2024-03-19T18:05:37.08+00:00

Hi! I'm getting the response "Distributing your web app across multiple instances" under the Diagnostics. I'm going through the steps for Autoscale under Monitor but I don't have the option to change it from Instance Count 1 to anything else, is it not available on the free plan perhaps?

Dan Rios 1,265 Reputation points MVP

2024-03-19T18:09:14.53+00:00

Those shouldn’t have any impact on the 504 errors you were seeing, as I see those often and they are just Microsoft recommendations for app availability. There should be an App crash or similar area where you may be able to see more specific information to your original issue? Is it still happening frequently? You can also go to the app console logs and check what that’s saying during the issue as well to get extra logs.

Dan Rios 1,265 Reputation points MVP

2024-03-19T18:13:11.8033333+00:00

As you’re on the free plan and getting 504 errors it may be due to maxing out compute power of the app. If you go to the overview of the app and review the monitoring metrics can you see memory or CPU maxing out? If so it may be that you need to consider scaling the plan to get more compute power if so.

Dio 0 Reputation points

2024-03-19T20:07:33.58+00:00

I don't see any errors about an app crash in the health diagnostics area. Previously the page would load after a long time of waiting like a few minutes length of time and now it seems to be working... would you happen to know what I can do to fix this in the future?

Dan Rios 1,265 Reputation points MVP

2024-03-19T20:10:00.95+00:00

I would review the apps CPU and memory metrics. As you’re on the free plan it may be that sometimes it is running out of compute capacity. You can review the performance metrics on the overview page of the Azure App and see if it is maxing out. If it has done during the issue then you may want to scale the plan to Basic and see if that helps solve your problem.

Dio 0 Reputation points

2024-03-20T02:38:29.76+00:00

I see! I'm reviewing the metrics and the overview of the app but nothing seems to jump out as overly used unless I'm looking at the wrong section if you meant another area?

Dan Rios 1,265 Reputation points MVP

2024-03-20T05:22:00.5266667+00:00

That should be it. You’ll have a CPU and memory graph where you can adjust the timeframe to see historical data. If you say, adjusted for last 7 days and it seems like the metrics look ok (no spikes to 100% for example), then maybe just keep an eye on this when the issue happens again to see if it is a compute limit thing. I would also check the performance metrics for the Azure SQL on your backend too on the overview to see if the issue may have been the backend too.
Sign in to comment