Autoinstrumentation and live telemetry

A financial services company has deployed a customer-facing web application for loan applications with Application Insights autoinstrumentation enabled. During peak business hours, the support team receives reports of intermittent slowdowns. The development team needs to correlate these user complaints with actual performance degradation in real time to determine if the issue is caused by increased traffic, external dependencies, or server resource constraints. They want to observe telemetry data with minimal latency (under five seconds) and see detailed breakdowns of server response codes, dependency call durations, and exception rates as they occur. Which approach provides the most effective solution for this investigation?