Which solution will meet these requirements?
Create an Amazon CloudWatch alarm that is based on the website’s logs that are published to a CloudWatch Logs log group. Configure the alarm to publish an SNS notification if the number of HTTP 4xx errors and 5xx errors exceeds a specified threshold.
Create an Amazon CloudWatch alarm that is based on the website’s published metrics in CloudWatch. Configure the alarm to publish an SNS notification that is based on anomaly detection.
Create an Amazon CloudWatch Synthetics heartbeat monitoring canary. Associate the canary with the website’s URL for end users. Create a CloudWatch alarm for the canary. Configure the alarm to publish an SNS notification if the value of the SuccessPercent metric is less than 99%.
Create an Amazon CloudWatch Synthetics broken link checker monitoring canary. Associate the canary with the website’s URL for end users. Create a CloudWatch alarm for the canary. Configure the alarm to publish an SNS notification if the value of the SuccessPercent metric is less than 99%.
Explanations:
This option monitors HTTP 4xx and 5xx errors, which indicate issues with user requests and server errors but does not directly measure overall website availability or user experience. It does not ensure that the uptime is monitored to be less than 99%.
While this option uses CloudWatch metrics and anomaly detection, it does not specifically address the website’s availability from the user’s perspective. It lacks a direct measurement of uptime that can accurately reflect user experience.
This option uses CloudWatch Synthetics to create a monitoring canary that simulates user traffic to the website. The SuccessPercent metric provides a clear indication of availability, and the alarm can notify via SNS if this value falls below 99%, which directly meets the requirement of monitoring uptime.
This option describes a broken link checker canary, which primarily focuses on checking for broken links rather than overall website availability. While it can provide some insights into user experience, it does not effectively measure the uptime percentage as required by the scenario.