AI Chat Template - Ollama resiliency modification should be in web app, not service defaults

[jongalloway]

Description

The latest AI Chat Template includes a modification for the resiliency when Ollama is selected, since Ollama calls can take a while (especially without good GPU support) and resiliency default is 10 seconds. This modification shouldn't be in the Extensions.cs in ServiceDefaults, since this will modify the HttpClient for all projects in the solution. It should just be in the Web project Program.cs.

Reproduction Steps

Create a new project with Ollama selected as the LLM.

Expected behavior

• ServiceDefaults Extensions.cs is the exact same as a new .NET Aspire project.
• Web project Program.cs includes the call to RemoveAllResilienceHandlers and add in standard resilience handler with extended timeouts.

Actual behavior

ServiceDefaults Extensions.cs is modified to affect all HttpClients in the entire solution:

#pragma warning disable EXTEXP0001 // RemoveAllResilienceHandlers is experimental
            http.RemoveAllResilienceHandlers();
#pragma warning restore EXTEXP0001

            // Turn on resilience by default
            http.AddStandardResilienceHandler(config =>
            {
                // Extend the HTTP Client timeout for Ollama
                config.AttemptTimeout.Timeout = TimeSpan.FromMinutes(3);

                // Must be at least double the AttemptTimeout to pass options validation
                config.CircuitBreaker.SamplingDuration = TimeSpan.FromMinutes(10);
                config.TotalRequestTimeout.Timeout = TimeSpan.FromMinutes(10);
            });

            // Turn on service discovery by default
            http.AddServiceDiscovery();
        });

Regression?

No, this was a new addition.

Known Workarounds

Delete code from ServiceDefaults and add it to Web project.

Configuration

No response

Other information

No response