AWS Certified Developer – Associate (DVA-C02) — Question 428

A developer is using an AWS Lambda function to process data. The developer needs to extract custom metrics about processing times from the Lambda logs. The developer needs to analyze the metrics, set alarms, and detect issues in real time.

Which solution will meet these requirements?

Answer options

• A. Publish custom metric data to AWS CloudTrail by using the PutMetricData API operation. Classify and collect the metrics. Create graphs and alarms in CloudTrail for the custom metrics.
• B. Use the open source client libraries provided by Amazon to generate the logs in the Amazon CloudWatch embedded metric format. Use CloudWatch to create the required graphs and alarms for the custom metrics.
• C. Use Amazon CloudWatch Logs Insights to create custom metrics by querying the logs that come from the Lambda function. Use CloudWatch to create the required graphs and alarms for the custom metrics.
• D. Create an Amazon Kinesis data stream to stream log events in real time from Lambda. Specify an Amazon S3 bucket as the destination for the Kinesis data stream. Use Amazon CloudWatch to visualize the log data and to set alarms.

Correct answer: B

Explanation

Using the CloudWatch embedded metric format (EMF) allows Lambda functions to output custom metrics asynchronously in log files, which CloudWatch then automatically extracts in real time for graphing and alarming. Option A is incorrect because CloudTrail is used for auditing API calls, not for custom performance metrics. Option C is incorrect because CloudWatch Logs Insights is a query tool for log analysis rather than a real-time metric extraction and alarming engine, and Option D introduces unnecessary latency and complexity by routing logs through Kinesis and S3.