MLA-C01 Practice Question: ML Solution Monitoring, Maintenance, and Security
A machine learning engineer notices that the latency of a SageMaker endpoint has increased over time. They need to identify which component (model inference vs. pre/post-processing) contributes most to the latency. Which CloudWatch metrics should they examine?
⚠ Common exam trap
It's easy for candidates to confuse the total Latency metric with a breakdown metric, assuming it alone can identify the bottleneck, when in fact only the pair of ModelLatency and OverheadLatency provides the necessary decomposition.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
ModelLatency and OverheadLatency
SageMaker endpoints emit CloudWatch metrics that break down total latency into model inference time (ModelLatency) and the time spent in pre/post-processing (OverheadLatency). By comparing these two metrics, the engineer can pinpoint whether the bottleneck is in the inference code or in the custom preprocessing/postprocessing logic. Option D directly provides both metrics needed for this root-cause analysis.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Latency and ModelLatency
Why it's wrong here
Latency is total end-to-end request time, so pairing it with ModelLatency only shows inference duration; the residual includes overheads but cannot isolate pre/post-processing specifically. It is tempting because both are latency metrics, and would be correct when you only need overall versus model-only response times.
- ✗
Invocations and 4XXError
Why it's wrong here
Invocations and 4XXError measure request volume and client-side errors, revealing nothing about where time is spent inside the container. They are tempting because invocation counts and error rates are standard endpoint health metrics, and would be correct when diagnosing traffic spikes or malformed client requests.
- ✗
5XXError and MemoryUtilization
Why it's wrong here
5XXError and MemoryUtilization show server faults and resource pressure, not the split between inference and pre/post-processing time. They are tempting because memory exhaustion genuinely inflates latency, and would be correct when diagnosing container crashes or out-of-memory conditions on the endpoint.
- ✓
ModelLatency and OverheadLatency
Why this is correct
ModelLatency isolates time spent in model inference, while OverheadLatency captures pre- and post-processing plus queueing outside the model. Comparing both CloudWatch metrics attributes the latency increase to the correct component, satisfying the stem's diagnostic requirement.
Go deeper
Related to this question
About these practice questions
Courseiva writes every MLA-C01 question from scratch — 665 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This MLA-C01 practice question is part of Courseiva's free Amazon Web Services certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the MLA-C01 exam.