X-Ray ML Tracing Quiz

5 questions Pass: 70% +25 pts (50 with Pro)

Quiz covering Troubleshooting ML

What you'll be asked

  1. What is the primary purpose of distributed tracing in an ML pipeline?
  2. In the context of X-Ray tracing for ML, what does a 'span' represent?
  3. You notice that your inference service is slow, but the model inference itself is fast. Which X-Ray feature would best help you identify the bottleneck?
  4. If you are debugging a production ML service and need to correlate a specific inference request to its corresponding trace, what is the best practice?
  5. When implementing custom instrumentation for an ML model inference container, why might you choose to wrap the 'predict' method in a subsegment rather than just the whole function?

Answers and explanations come after you submit.