Refines timeout and network error handling

[magicxor]

Improves the application's resilience to network interruptions and slow API responses:

• Reduces Google Cloud Request Timeout: Sets a shorter default timeout for HTTP client requests to Google Cloud services, allowing for quicker detection of unresponsive calls.
• Enhances Chat Service Error Handling:
  • Expands the retry policy to cover various timeout and network I/O exceptions (TimeoutRejectedException, TimeoutException, HttpIOException, TaskCanceledException with inner TimeoutException).
  • Introduces dedicated error handling to automatically switch to a fallback chat model when these errors occur, improving user experience during service disruptions.
• Broadens Image Generation Service Error Detection: Updates the IsModelUnavailableException logic to include a wider range of HTTP, network I/O, and timeout exceptions, leading to more robust error classification for generative AI model availability.

[Copilot]

Pull request overview

Improves resilience when calling Google Generative AI / Google Cloud services by tightening request timeouts and expanding transient error detection, including switching chat requests to a fallback model after certain failures.

Changes:

• Reduced the configured Google Cloud HttpClient timeout to 40 seconds.
• Expanded retry conditions in MentionHandler to include additional timeout/network exception types and added fallback switching on those failures.
• Broadened image model “unavailable” exception detection to include additional HTTP/network/timeout exception types.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
Pacos/Services/GenerativeAi/ImageGenerationService.cs	Expands the set of exceptions treated as “model unavailable” to trigger retry/fallback attempts.
Pacos/Services/ChatCommandHandlers/MentionHandler.cs	Adds more retryable exception types and introduces fallback switching for timeout/network IO failures.
Pacos/Program.cs	Shortens the Google Cloud HttpClient timeout configuration to detect stalled calls sooner.

Comments suppressed due to low confidence (1)

Pacos/Program.cs:72

• With AddStandardResilienceHandler configuring AttemptTimeout and TotalRequestTimeout, setting HttpClient.Timeout to googleRequestTimeout can cap the entire SendAsync operation (including retries) at 40s, preventing TotalRequestTimeout (and any built-in retries) from being effective. Consider setting HttpClient.Timeout to Timeout.InfiniteTimeSpan (or at least >= TotalRequestTimeout) and relying on the resilience handler timeouts for enforcement.

                    var googleRequestTimeout = TimeSpan.FromSeconds(40);
                    services.AddHttpClient(nameof(HttpClientType.GoogleCloud), httpClient => httpClient.Timeout = googleRequestTimeout)
                        .ConfigurePrimaryHttpMessageHandler((handler, serviceProvider) =>
                        {
                            var proxyAddress = serviceProvider.GetRequiredService<IOptions<PacosOptions>>().Value.WebProxy;

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.