You monitor an Azure web app with Application Insights and must trigger an alert when the application experiences a sudden increase in performance issues and failures. Which feature should you use?

Smart Detection automatically analyzes telemetry from your application and proactively warns you about potential performance problems and failures. It uses machine learning to detect anomalies like sudden increases in failure rates, performance degradation, and memory leaks, making it ideal for triggering alerts on unexpected issues. Custom events are user-defined events for tracking specific actions, not for automatic anomaly detection. Application Insights Profiler is used for in-depth performance analysis of code paths, not for real-time anomaly alerting. Usage analysis focuses on user behavior and application adoption. Continuous export sends telemetry to other storage locations but doesn't provide anomaly detection or alerting itself.