[Bug] Memory usage when upload file to DataLake gen2

[Marusyk]

• List of NuGet packages and version that you are using:
  Microsoft.ApplicationInsights.AspNetCore 2.14.0
  Microsoft.ApplicationInsights.WorkerService 2.14.0
• Runtime version (e.g. net461, net48, netcoreapp2.1, netcoreapp3.1, etc. You can find this information from the *.csproj file): netcoreapp3.1
• Hosting environment (e.g. Azure Web App, App Service on Linux, Windows, Ubuntu, etc.): Linux Docker

I'm using Azure.Storage.Files.DataLake to send data every minute from my C# code with enabled Application Insights

services.AddApplicationInsightsTelemetryWorkerService();

There are a lot of operations with Datalake (now it's more than 5000 per hour).
I got a lot of memory usage of my application (each file ~150 bytes).

App without Application Insights enabled, after 3 days uptime: ~96Mb
App with Application Insights disabled, after 6 hours uptime: ~900Mb

Could you please suggest how to avoid AI memory leak?

I've tried but no success

public class DependencyTelemetryFilter : ITelemetryProcessor
    {
        private ITelemetryProcessor Next { get; set; }

        public DependencyTelemetryFilter(ITelemetryProcessor next)
        {
            Next = next;
        }

        public void Process(ITelemetry item)
        {
            if (item is DependencyTelemetry elem && elem.Contains("dfs.core.windows.net"))
            {
                return;
            }

            Next.Process(item);
        }
    }

[Marusyk]

/ping

[cijothomas]

@Marusyk Could you share the code snippet used to make datalake calls? This issue can occur if the azure-data-lake sdk is used incorrectly as well, so it'd be great if you can share a small repro or code snippet doing datalake calls.

[Marusyk]

I'm using Azure.Storage.Files.DataLake to send data every minute from my C# code.

public class DataLakeStorage : IFileStorage
{
	private readonly DataLakeServiceClient _dataLakeClient;

	public DataLakeStorage(FileStorageOptions options)
	{
		var sharedKeyCredentials = new StorageSharedKeyCredential(options.AccountName, options.AccountKey);
		_dataLakeClient = new DataLakeServiceClient(new Uri($"https://{options.AccountName}.dfs.core.windows.net"), sharedKeyCredentials);
	}

	public async Task Add(string containerName, string path, string name, object data)
	{
		DataLakeFileSystemClient fileSystemClient = _dataLakeClient.GetFileSystemClient(containerName.ToLower());
		await fileSystemClient.CreateIfNotExistsAsync();

		DataLakeDirectoryClient directoryClient = fileSystemClient.GetDirectoryClient(path);
		DataLakeFileClient fileClient = await directoryClient.CreateFileAsync(name);

		await using var stream = new MemoryStream();
                await JsonSerializer.SerializeAsync(stream, data, data.GetType());
                stream.Position = 0;
                await fileClient.AppendAsync(stream, 0);
                await fileClient.FlushAsync(stream.Length);
	}
}

The DataLakeStorage is registered as a Singleton in web application.
Also, tried await fileClient.UploadAsync(stream, overwrite: true); instaed of

await fileClient.AppendAsync(stream, 0);
await fileClient.FlushAsync(stream.Length);

and added

 <PropertyGroup>
    <ServerGarbageCollection>false</ServerGarbageCollection>
  </PropertyGroup>

but the same result. It uses a lot of memory.

[cijothomas]

@Marusyk The code does not appear to be concerning. I wanted to first see if this is related to #1867.
As this is not related, could you share a repro project for us to do deeper investigation?

[Marusyk]

@cijothomas
Actually, you need only this code with
services.AddApplicationInsightsTelemetryWorkerService(); to reprocude it.

call every few seconds:

await _instance.Add("containerName", DateTime.Now.ToString(), "name", new { Name = "name", Time = DateTime.Now });

[Marusyk]

any updates?
I upgraded my application to .NET 5 but it didn't help

[wesselkranenborg]

We are also facing this issue. We now implemented a regular restart of our service but that is far from ideal.

[Marusyk]

Hi @cijothomas, can anyone help with it?
It is not funny now. We got a huge memory usage when AppInsights enabled!!!

[github-actions]

This issue is stale because it has been open 300 days with no activity. Remove stale label or comment or this will be closed in 7 days.

[Marusyk]

The issue is not fixed