Free AWS AIP-C01 Actual Exam Questions - Question 13 Discussion
Scenario: A recommendation endpoint experiences significant delays during predictable high-traffic sales events, resulting in poor user experience. The goal is to adjust the target tracking scaling policy to proactively ensure sufficient capacity and prevent latency issues during these peak periods. Question- Which solution will best meet the requirements?. Options:
I’m ruling out A since restarting the endpoint won’t really add capacity or reduce latency during peak times. Between B and D, B seems better because it’s proactive and planned around known traffic spikes, which fits the scenario of predictable sales events. D reacts to utilization but might lag behind when demand suddenly jumps, causing delays. C could help but doesn’t scale dynamically and might be costly if the big instance is idle most times. Does the question hint at how precise the traffic patterns are for scheduled scaling to be reliable?
B/D? B makes sense since it’s proactive, ensuring capacity is ready before the spike. But D offers flexibility to respond if the traffic varies unexpectedly during the event, scaling based on actual resource use. I’d pick B for guaranteed prep, but D could help avoid wasted resources if the load isn’t as high as predicted. A and C seem less targeted—A doesn’t add capacity, and C might be costly without addressing timing and scaling specifics.
I don’t think restarting with Lambda (A) solves the main issue since it doesn’t add capacity. Option B is solid because it scales proactively before the event starts, which is critical here. B
Maybe D could work here too since step scaling adjusts resources based on actual load signals like CPU or memory. It’s more dynamic than just a scheduled bump and might prevent over-provisioning if the traffic isn’t as high as expected. Still, it might react a bit slower than scheduled scaling during sudden spikes. But it’s worth considering if you want the system to adapt continuously rather than just relying on a fixed schedule.
It’s B for sure. Scheduled scaling lets you ramp up capacity right before the event starts, so you’re not playing catch-up during the spike. D is more reactive—by the time CPU or memory usage triggers scaling, users are already experiencing delays. Also, A and C don’t really address the timing issue; restarting or just bigger instances won’t help if the endpoint isn’t ready before demand hits. So proactive scaling with a schedule is the way to go here.
Option B makes sense; proactive scheduled scaling is key before traffic spikes.
It’s B because scheduled scaling avoids delays by adding capacity ahead of time.
B, since scheduled scaling ensures capacity is ready before the spike hits.
Makes sense to go with D since it’s fully managed and auto-clusters topics, no manual labeling or training needed. That’s way less hassle than options needing custom classifiers or data prep. D
D feels right because it uses Comprehend’s ready-made topic detection, so no need to build or train anything custom. It’s straightforward and fits the “minimize effort” part well.
C imo makes sense too since it uses a custom classifier to assign predefined labels, which could be more precise than unsupervised topic detection in D. It still avoids heavy custom model building.
I’m leaning towards D because it uses Amazon Comprehend’s built-in topic detection, which seems like the least custom effort compared to building classifiers or training models. It directly clusters topics without extra manual labeling. Anyone else think this is the simplest way?