What should the solutions architect do to meet this requirement?
Send Amazon CloudWatch logs to Amazon Redshift. Use Amazon QuickS ght to perform further analysis.
Enable detailed monitoring on all EC2 instances. Use Amazon CloudWatch metrics to perform further analysis.
Create an AWS Lambda function to fetch EC2 logs from Amazon CloudWatch Logs. Use Amazon CloudWatch metrics to perform further analysis.
Send EC2 logs to Amazon S3. Use Amazon Redshift to fetch logs from the S3 bucket to process raw data for further analysis with Amazon QuickSight.
Explanations:
Sending Amazon CloudWatch logs to Amazon Redshift for analysis does not provide the required performance granularity of no more than 2 minutes, as Redshift is primarily for analytics rather than real-time monitoring. QuickSight is also not designed for immediate analysis of CloudWatch logs.
Enabling detailed monitoring on all EC2 instances provides metrics at a granularity of 1 minute (the lowest granularity available), which allows for real-time performance monitoring of the application. This meets the requirement of analyzing performance with no more than 2 minutes of granularity.
While fetching EC2 logs from CloudWatch Logs using a Lambda function can provide some log data, it does not directly address the requirement for performance analysis with a granularity of 2 minutes. Additionally, CloudWatch metrics provide a more straightforward solution for performance analysis.
Sending EC2 logs to S3 and using Redshift for analysis introduces unnecessary complexity and does not provide the required performance granularity of 2 minutes. The process of fetching and analyzing logs from S3 adds latency and does not directly allow for timely performance monitoring.
I think the answer is:
Enable detailed monitoring on all EC2 instances. Use Amazon CloudWatch metrics to perform further analysis.