DEA-C01 Practice Question 31: Kinesis on Demand Capacity Mode
AWS Certified Data Engineer – Associate (DEA-C01) · Test 4 · question 1 of 10
A retailer's clickstream flows into an Amazon Kinesis data stream in provisioned mode with 20 shards. Traffic is flat most of the year but multiplies during flash sales, when producers receive ProvisionedThroughputExceededException errors. Outside those events the shards sit at roughly 10 percent utilization. Nobody on the team wants to own shard-count planning.
Which change addresses both the throttling and the idle capacity?
Show answer and explanation
Answer:
On-demand streams "require no capacity planning and automatically scale to handle gigabytes of write and read throughput per minute", and the documentation states plainly that on-demand mode "is ideal for addressing the needs of highly variable and unpredictable application traffic. You no longer have to provision these workloads for peak capacity, which can result in higher costs due to low utilization." Billing moves to a per-GB model, so the quiet months stop paying for idle shards. A stream in on-demand mode accommodates up to double the peak write throughput seen in the previous 30 days and scales as new peaks arrive. You can switch modes twice in 24 hours with no disruption to producers or consumers.
Option A fixes the throttling by paying for peak capacity all year, exactly the low-utilization cost the team is already unhappy about, made worse.
Option B is on-demand mode rebuilt by hand, and it reacts an hour late. A flash sale that starts at the top of the hour throttles for the whole hour before the function notices. UpdateShardCount also has scaling limits and takes time to complete, so it cannot chase a sudden spike.
Option D pushes sharding into the producers. It doubles the streams to monitor, splits the consumer story in two, and still requires someone to decide how many shards each stream needs, no part of the original problem is solved.
Reference: Choose the right mode to stream in