Send Request Times过高原因排查及AppInsights相关性能疑问

First, let’s break this down: the Send Request Time tracked by App Insights for AJAX calls only covers that specific request’s network transit and server processing, but your total page load time includes far more than just AJAX calls. Here are the most likely culprits behind the 4.27s total load:

• Initial HTML Request Overhead: You might be focusing on AJAX calls, but the initial request to fetch your page’s core HTML could be the bottleneck. Check the Page View Telemetry in App Insights for the initial page load’s duration and sendRequestTime—this often includes DNS lookup, TCP handshake, and server-side rendering time that isn’t captured in your AJAX metrics.
• Unoptimized Static Resources: CSS, JavaScript, images, and fonts add up quickly. If these assets are large, uncompressed, or not cached properly, they’ll drag down total load time. Use App Insights’ Browser Timings panel to look at metrics like networkDuration for static assets—you might find a huge unoptimized image or unminified JS file taking seconds to load.
• Render-Blocking Resources: If your critical CSS is loaded externally (not inlined) or you’re loading synchronous JS early in the page, the browser will pause rendering until those resources are processed. This time doesn’t count toward AJAX request times but is included in total page load. Use your browser’s dev tools Performance tab to spot these blocking resources.
• Resource Loading Order: If your page loads resources sequentially instead of in parallel (e.g., forgetting to use async/defer for non-critical JS), the cumulative wait time adds up. Even if each individual request is fast, waiting for one to finish before starting another can balloon the total load time.
• Third-Party Scripts: Tools like analytics, ads, or chat widgets often load asynchronously but can still block rendering or add hidden network overhead. Check your browser’s Network tab for third-party requests—these are easy to miss but can contribute significantly to total load time.

It’s strange that one site in your Web App shows server response charts while the other doesn’t. Here’s what to investigate:

• Server-Side SDK Configuration: Double-check that the problematic site has the App Insights server-side SDK properly installed and configured. For example, if using ASP.NET, verify the Microsoft.ApplicationInsights.Web NuGet package is installed, and your ApplicationInsights.config has the correct Instrumentation Key and includes the ServerTelemetryModule.
• Telemetry Sampling or Filtering: Check if sampling is enabled for the missing site—if the sampling rate is set too low, server-side telemetry might not be captured. Also, look for custom telemetry processors that could be filtering out server response data. You can adjust sampling settings in ApplicationInsights.config or via the Azure Portal’s App Insights settings.
• Deployment Environment Differences: If the two sites are deployed to different slots or have distinct app settings, ensure the missing site has the App Insights extension enabled (for Azure App Service) or that IIS is configured to collect server telemetry. Staging slots or custom deployment scripts sometimes disable telemetry by accident.
• Permission Issues: Confirm your Azure account has the correct permissions to view server-side telemetry for the problematic site. If using RBAC, you might need the Monitoring Reader or Contributor role for that specific App Insights resource.
• Telemetry Delivery Failures: Check the Telemetry Status panel in App Insights for the missing site—this will show if server-side telemetry is failing to send (e.g., due to network issues or authentication problems). You can also review server logs for errors related to App Insights telemetry submission.

内容的提问来源于stack exchange，提问作者Pio