AWS Certified Solutions Architect – Professional — Question 410

A company is building a sensor data collection pipeline in which thousands of sensors write data to an Amazon Simple Queue Service (Amazon SQS) queue every minute. The queue is processed by an AWS Lambda function that extracts a standard set of metrics from the sensor data. The company wants to send the data to Amazon CloudWatch. The solution should allow for viewing individual and aggregate sensor metrics and interactively querying the sensor log data using
CloudWatch Logs Insights.
What is the MOST cost-effective solution that meets these requirements?

Answer options

• A. Write the processed data to CloudWatch Logs in the CloudWatch embedded metric format.
• B. Write the processed data to CloudWatch Logs. Then write the data to CloudWatch by using the PutMetricData API call.
• C. Write the processed data to CloudWatch Logs in a structured format. Create a CloudWatch metric filter to parse the logs and publish the metrics to CloudWatch with dimensions to uniquely identify a sensor.
• D. Configure the CloudWatch Logs agent for AWS Lambda. Output the metrics for each sensor in statsd format with tags to uniquely identify a sensor. Write the processed data to CloudWatch Logs.

Correct answer: C

Explanation

Option C is the most cost-effective solution because CloudWatch metric filters allow you to extract metric observations from ingested log events at no extra cost, avoiding the expensive API charges associated with PutMetricData (Option B). This approach satisfies the interactive querying requirement via CloudWatch Logs Insights while maintaining individual and aggregate views through dimensions. Using the CloudWatch embedded metric format (Option A) or custom agents (Option D) would generate higher operational overhead and potentially higher costs due to the high cardinality of thousands of sensors.