A company hosts a multi-tier public web application on Amazon EC2 with an RDS database. The company expects a large increase in sales over a holiday weekend. A solutions architect must build a solution to analyze web application performance with a granularity of no more than 2 minutes. What should the solutions architect do?

Enabling detailed monitoring on EC2 instances provides metrics at 1-minute granularity, which satisfies the requirement of "no more than 2 minutes." These metrics are directly available in Amazon CloudWatch, allowing for immediate analysis of web application performance, including CPU utilization, network I/O, and disk activity. CloudWatch metrics are designed for this type of performance analysis. Incorrect options: Sending CloudWatch Logs to Redshift and using QuickSight is for log analysis and business intelligence, not real-time performance metrics. Using a Lambda function to fetch EC2 logs from CloudWatch Logs is an overly complex and indirect method for performance analysis when direct metrics are available. Logs provide event data, not aggregated performance metrics. Sending EC2 logs to S3 and then Redshift for QuickSight analysis is also focused on log analysis and data warehousing, not direct, granular performance monitoring. This approach introduces significant latency and complexity for the stated goal.