Distill
Prep
Python
GenAI
GenAI Frameworks
NLP
Deep Learning
Machine Learning
ML Libraries
Statistics
SQL
MLOps
Cloud
System Design
Pricing
Blog
Learn
Practice
Test
Live Engine
Select Topic
Easy
(33)
Medium
(33)
Hard
(33)
easy
Aws Sagemaker
A team deploys a SageMaker real-time endpoint with one instance. Traffic is low on weekends (5 RPS) and high on weekdays (80 RPS). What is the simplest AWS-native solution to automatically handle this traffic difference without overpaying?
A
Deploy two separate endpoints — one for weekdays, one for weekends — and update DNS to switch between them
B
Enable Application Auto Scaling on the SageMaker endpoint with a scaling policy based on
InvocationsPerInstance
metric. Set
MinCapacity=1
(handles weekends) and
MaxCapacity=4
(handles weekday peaks). The endpoint scales out as traffic increases and scales in during low periods
C
SageMaker endpoints cannot scale; provision for peak traffic permanently
D
Use a scheduled Lambda function to manually update the endpoint's instance count at 9am Monday and 5pm Friday
Confirm Answer
Previous
Back
Next