.Net: The operation was cancelled because it exceeded the configured timeout of 0:01:40

[MarkusMeyer13]

Describe the bug

I'm using Microsoft.SemanticKernel version="1.0.0-beta2"
The InvokeAsync method raises an exception:

The operation was cancelled because it exceeded the configured timeout of 0:01:40. Network timeout can be adjusted in ClientOptions.Retry.NetworkTimeout.

StackTrace:

at Azure.Core.Pipeline.ResponseBodyPolicy.ThrowIfCancellationRequestedOrTimeout(CancellationToken originalToken, CancellationToken timeoutToken, Exception inner, TimeSpan timeout)
at Azure.Core.Pipeline.ResponseBodyPolicy.d__5.MoveNext()
at Azure.Core.Pipeline.RedirectPolicy.d__7.MoveNext()
at Azure.Core.Pipeline.RetryPolicy.d__5.MoveNext()
at Azure.Core.Pipeline.RetryPolicy.d__5.MoveNext()
at Azure.Core.HttpPipelineExtensions.d__0.MoveNext()
at System.Threading.Tasks.ValueTask1.get_Result() at Azure.AI.OpenAI.OpenAIClient.<GetChatCompletionsAsync>d__17.MoveNext() at Microsoft.SemanticKernel.Connectors.AI.OpenAI.AzureSdk.ClientBase.<RunRequestAsync>d__291.MoveNext()
at Microsoft.SemanticKernel.Connectors.AI.OpenAI.AzureSdk.ClientBase.d__19.MoveNext()
at Microsoft.SemanticKernel.Connectors.AI.OpenAI.AzureSdk.ClientBase.d__22.MoveNext()
at Microsoft.SemanticKernel.SemanticFunction.d__36.MoveNext()
at Microsoft.SemanticKernel.SemanticFunction.d__17.MoveNext()

To Reproduce
Steps to reproduce the behavior:

This is the C# code which produces the exception:

static async Task<string?> Summarize(string openAiDeploymentName, string openAiEndPoint, string openAiKey, string prompt, string data)
{
    try
    {
        var builder = new KernelBuilder();
        builder.WithAzureChatCompletionService(openAiDeploymentName, openAiEndPoint, openAiKey);
        var kernel = builder.Build();

        var summarize = kernel.CreateSemanticFunction(prompt, requestSettings: new OpenAIRequestSettings() { Temperature = 0.1 });

        // 'data' has a length of 8993
        var result = await summarize.InvokeAsync(data, kernel).ConfigureAwait(false);

        return result.GetValue<string>();
    }
    catch (Exception e)
    {
        await Console.Out.WriteLineAsync(e.Message);
        throw;
    }
}

Used model:

Model name: gpt-35-turbo-16k
Model version: 0613
Version update policy: Once a new default version is available.
Deployment type: Standard
Content Filter: Default
Tokens per Minute Rate Limit (thousands): 5
Rate limit (Tokens per minute): 5000
Rate limit (Requests per minute): 60

The data to summarize has a length of 8993 characters.
The timeout does not happen if the length is e.g. 4000 characters.

1. Run code
2. Wait for the result
3. TimeOut exception occurs

Expected behavior
InvokeAsync shall return a valid result also for longer strings.

Screenshots

Platform

• OS: Windows
• IDE: Visual Studio
• Language: C#

Additional context
Add any other context about the problem here.

[matthewbolanos]

@MarkusMeyer13, does providing a custom HTTP client to the kernel with a different timeout fix the issue?

[MarkusMeyer13]

I solved the timeout issue by increasing the timeout value with OpenAIClientOptions:

var clientOptions = new OpenAIClientOptions();
clientOptions.Retry.MaxRetries = 1;
clientOptions.Retry.NetworkTimeout = TimeSpan.FromMinutes(4);
var openAIClient = new OpenAIClient(new Uri(openAiEndPoint), new AzureKeyCredential(openAiKey), clientOptions);

var builder = new KernelBuilder();
builder.WithAzureChatCompletionService(openAiDeploymentName, openAIClient);
var kernel = builder.Build();