PMLE Monitoring ML Solutions Practice Question

An ML engineer needs to monitor the online prediction latency of a Vertex AI Endpoint. Which metrics should they look at in Cloud Monitoring?

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

✓

p50, p95, p99 latency

Cloud Monitoring provides latency metrics for Vertex AI Endpoints, including p50, p95, and p99 latency, which are key for understanding performance.

Answer analysis

For each option: why learners choose it and why it is or isn't the right answer here.

✓
p50, p95, p99 latency
Why this is correct
Correct: These percentiles are standard for monitoring prediction latency.
✗
Request count and error rate
Why it's wrong here
These are important but do not directly measure latency.
✗
Skew and drift scores
Why it's wrong here
Skew and drift are data distribution metrics, not latency.
✗
CPU/GPU utilization
Why it's wrong here
CPU/GPU utilization monitors resource usage but not prediction latency.

Go deeper