A telco collects several thousand network usage data points per second and processes them in real time, aggregating and storing results in an Amazon Aurora DB instance. Sudden falls in usage typically indicate outages, and the company needs to detect those drops and respond immediately with the lowest possible latency. Which solution provides the least-latency detection?

The correct solution uses Amazon Kinesis Data Streams and Amazon Managed Service for Apache Flink because this combination is designed for real-time, low-latency processing of high-volume streaming data. Kinesis Data Streams ingests the data as it arrives, and Flink processes it continuously, allowing for immediate detection of usage drops. The incorrect options introduce latency. Querying Aurora or DynamoDB every minute with Lambda creates a significant delay, as the data is not processed in real time. While DynamoDB and DAX offer low-latency access, the minute-long polling interval remains a bottleneck. Aurora Database Activity Streams are for monitoring database activity, not for real-time analysis of application-level metrics like network usage drops.