Containerized ASP.NET app on Azure App Service F1 warmup times out on first run

Question

Containerized ASP.NET app on Azure App Service F1 warmup times out on first run

Jan Hein de Jong 0

After having done mostly on-prem stuff at my company, we decided we want to give Azure a try. I'm trying to run a containerized ASP.NET app (with react frontend served as static files), on a free-tier (F1) app service.

When I haven't called the app in a while - as expected - the service goes to sleep and kills the container. When a client tries to connect with my app, the service tries to relaunch the container, by pulling the container, starting it (and some adjacent MSI container), and starts checking readiness by sending warmup requests. So far, so good.

Now the weird stuff starts.

These warmup requests time-out after 230 s, but if I inspect the container logs, I see the app is already up and listening. After the warmup timeout, the app service kills the container, and repeats the process of pulling, starting, and sending warmup requests, which now goes through almost instantly (and my app works fine).

Any ideas what could cause this behavior? A quick fix would be to set this warmup timeout to something like 2-3 seconds, but it feels weird.

SnehaAgrawal-MSFT 22,721 Reputation points Moderator

2024-03-01T06:33:54.4466667+00:00
@Jan Hein de Jong Thanks for reaching here!

It seems like you are encountering an issue with the warmup process of your containerized ASP.NET app on Azure App Service F1. The warmup requests are timing out after 230 seconds, even though the app is already up and listening according to the container logs. This is causing the app service to kill the container and repeat the process of pulling, starting, and sending warmup requests.

One possible cause of this behavior could be the limited resources available on the F1 tier which may not be sufficient for your app to start up quickly. When the app service sends warmup requests, it may be overwhelming the available resources, causing the requests to time out.

To troubleshoot this issue, you can try the following steps:

Try increasing the warmup timeout to a higher value to see if that resolves the issue. You can do this by setting the WEBSITES_CONTAINER_START_TIME_LIMIT app setting to a higher value. To configure that, add an app setting called WEBSITES_CONTAINER_START_TIME_LIMIT and set it to the number of seconds you would like for us to wait for your container to start (up to a maximum of 1800) as shown in the image below.

You may also try scaling up to a higher pricing tier to see if the issue persists. If the issue goes away when you scale up, it's possible that the limited resources are causing issues with the warmup process.

Further suggest you to refer this detailed blog on- Things You Should Know: Web Apps and Linux

I hope this helps you resolve the issue. If you have any further questions or concerns, please let me know.
Jan Hein de Jong 0 Reputation points

2024-03-01T09:24:04.69+00:00

Yes, I had already tried that.

When I set the start time to - say - 10 seconds, the same happens. First time the warmup request cycle times out (after 10 seconds this time, instead of 230). The container is pulled and started again, and this time it works almost instantly.

Since my app is a .NET app, I've switched from using containers, to deploying the app directly on the App Service. Now it works like a charm.
Jaap van der Minne 0 Reputation points

2024-05-07T10:20:04.91+00:00

I have exactly the same issue with some Python apps. Did you ever find a fix for this?

Your answer

SnehaAgrawal-MSFT 22,721 Reputation points Moderator

2024-03-01T06:33:54.4466667+00:00

@Jan Hein de Jong Thanks for reaching here!

It seems like you are encountering an issue with the warmup process of your containerized ASP.NET app on Azure App Service F1. The warmup requests are timing out after 230 seconds, even though the app is already up and listening according to the container logs. This is causing the app service to kill the container and repeat the process of pulling, starting, and sending warmup requests.

One possible cause of this behavior could be the limited resources available on the F1 tier which may not be sufficient for your app to start up quickly. When the app service sends warmup requests, it may be overwhelming the available resources, causing the requests to time out.

To troubleshoot this issue, you can try the following steps:

Try increasing the warmup timeout to a higher value to see if that resolves the issue. You can do this by setting the WEBSITES_CONTAINER_START_TIME_LIMIT app setting to a higher value. To configure that, add an app setting called WEBSITES_CONTAINER_START_TIME_LIMIT and set it to the number of seconds you would like for us to wait for your container to start (up to a maximum of 1800) as shown in the image below.

You may also try scaling up to a higher pricing tier to see if the issue persists. If the issue goes away when you scale up, it's possible that the limited resources are causing issues with the warmup process.

Further suggest you to refer this detailed blog on- Things You Should Know: Web Apps and Linux

I hope this helps you resolve the issue. If you have any further questions or concerns, please let me know.
Jan Hein de Jong 0 Reputation points

2024-03-01T09:24:04.69+00:00

Yes, I had already tried that.

When I set the start time to - say - 10 seconds, the same happens. First time the warmup request cycle times out (after 10 seconds this time, instead of 230). The container is pulled and started again, and this time it works almost instantly.

Since my app is a .NET app, I've switched from using containers, to deploying the app directly on the App Service. Now it works like a charm.
Jaap van der Minne 0 Reputation points

2024-05-07T10:20:04.91+00:00

I have exactly the same issue with some Python apps. Did you ever find a fix for this?

Share via

Containerized ASP.NET app on Azure App Service F1 warmup times out on first run

Your answer