Azure function intermittent error during cold start after converting to .NET 8 isolated worker

Question

I have an Http Trigger Azure function that has been working fine for the past 2 years. It is an Azure Function running on a Windows Consumption App Service Plan. I recently updated from .net6 in-process to .net8 isolated worker model. After the upgrade to isolated worker, I am now seeing intermittent errors when calling the function. The function seems to crash after a cold start and errors are written to AppInsights. Calling the function a second time a few seconds later (typically 15 seconds) - the call will succeed and then subsequent calls will generally work for a while. Then we will typically see another error after the next cold start - although not always - sometimes it works fine.

This problem did not exist before migrating to the isolated worker. I have upgraded to all the latest NuGet packages but still the problems persist. I have seen other people posting similar issues after upgrading to the isolated worker model but I've not managed to find a resolution for this as yet.

I've got a lot of Azure Functions running on Linux Consumption Plans and have not seen this error on any of them - just this Windows one.

The stack trace includes a number of exceptions with the following error messages:

• "Timed out waiting for the function start call. Invocation: 'f5750d96-7cf5-4d66-86ff-9a229e497006'."
• An attempt was made to access a socket in a way forbidden by its access permissions.
• Failed to proxy request with ForwarderError
• Exception while executing function: Functions.CreatePdfDocument
• System.Net.Sockets.SocketException at Yarp.ReverseProxy.Forwarder.HttpForwarder+

Answer 1

I had the same issue. Found this article https://stackoverflow.com/questions/76952125/isolated-function-app-not-responding-to-http-calls

In my case, I commented out one item in host.json, which started to cause the issue. host.json cannot have comments (starting with some version of Azure Function)

Answer 2

Just for reference here's my host.json. For the last 12 hours, I only had 3 traces in my App Insights logs, all valid warnings.

{
  "version": "2.0",
  "logging": {
    "applicationInsights": {
      "samplingSettings": {
        "isEnabled": true,
        "maxTelemetryItemsPerSecond": 20,
        "excludedTypes": "Request;Exception"
      },
      "enableLiveMetricsFilters": true,
      "enableDependencyTracking": true
    },
    "logLevel": {
      "default": "Information",
      "Function": "Information",
      "Host.Aggregator": "Warning",
      "Azure.Core": "Warning",
      "Azure.Storage.Queues": "Warning",
      "Azure.Storage.Blobs": "Warning",
      "Host.Controllers.Host": "Warning",
      "Microsoft.AspNetCore.Mvc.StatusCodeResult": "Warning",
      "Microsoft.AspNetCore.Mvc.ChallengeResult": "Warning",
      "Microsoft.Azure.WebJobs.Hosting.OptionsLoggingService": "Warning"
    }
  }
}