MLA-C01 Question 407
Select 3You are a Machine Learning Engineer managing a real-time inference system deployed on AWS Lambda. Recently, users have reported increased latency in predictions. You need to identify the root cause of the issue using AWS monitoring and observability tools. Which combination of actions should you take to troubleshoot the performance problem?
- A
Use AWS X-Ray to trace the end-to-end request flow and identify bottlenecks in the system.
- B
Enable Amazon CloudWatch Lambda Insights to monitor the function’s CPU, memory usage, and execution duration.
- C
Query Amazon CloudWatch Logs Insights to analyze and filter logs for error patterns and warnings.
- D
Use Amazon SageMaker Model Monitor to inspect the quality of the model's predictions.
- E
Enable Amazon S3 Event Notifications to track changes in the S3 bucket storing inference results.
Show answer and explanation
Correct answers: A, B, C
Explanation
To troubleshoot performance issues in a Lambda-based inference system, you need to leverage observability tools that provide insights into request flow, resource utilization, and log analysis. AWS X-Ray helps trace requests and identify bottlenecks, CloudWatch Lambda Insights provides detailed function-level metrics, and CloudWatch Logs Insights enables you to extract meaningful information from logs. These tools together offer a comprehensive approach to identify and resolve latency issues effectively.
- A. Correct.
AWS X-Ray provides end-to-end request tracing, which helps identify latency bottlenecks across the system, including Lambda functions, databases, and external services.
- B. Correct.
Amazon CloudWatch Lambda Insights provides detailed performance metrics such as CPU utilization, memory usage, and execution duration, which are critical for diagnosing performance issues in Lambda functions.
- C. Correct.
Amazon CloudWatch Logs Insights allows you to query and analyze log data, which can help identify specific errors or anomalies contributing to performance issues.
- D. Incorrect.
Amazon SageMaker Model Monitor is used to detect data quality issues or drift in SageMaker models, but it is not directly applicable to troubleshooting Lambda inference latency.
- E. Incorrect.
Amazon S3 Event Notifications is used for tracking changes in S3 buckets, but it is unrelated to diagnosing performance issues in a Lambda-based inference system.