A media outlet runs a public news and video site on AWS. The site keeps an index of video metadata in an Amazon DynamoDB table configured with provisioned capacity, and the video files themselves reside in an Amazon S3 bucket. During a high-traffic event, reads against the DynamoDB table were throttled and videos failed to appear. Operators temporarily raised the table's provisioned read capacity manually. The team wants to be alerted before throttling occurs in the future. An Amazon SNS topic has been created and the operators' email is subscribed. What action should the team take next to meet this requirement?

The correct option is to create a CloudWatch alarm on the ConsumedReadCapacityUnits metric. This metric directly reflects the actual read throughput consumed by the DynamoDB table. By setting the alarm threshold close to the provisioned read capacity, the team will be proactively notified when read consumption approaches the limit, allowing them to intervene before throttling occurs. Enabling DynamoDB auto scaling would prevent throttling by automatically adjusting capacity, but it doesn't directly address the requirement to be alerted before throttling. EventBridge rules for scaling actions would notify after scaling, not before throttling. Monitoring THROTTLINGEXCEPTION status codes in CloudWatch Logs or application logs would only alert after throttling has already happened, which is too late to prevent the issue. The goal is to be alerted before throttling.