passdrill
AWS SAA-C03 · RDS, DynamoDB & Databases · Card 021/024 medium

A retailer's Aurora database experiences heavy load only during periodic flash sales and is otherwise nearly idle, and the team wants compute capacity to scale up and down automatically in fine-grained increments within seconds, without manually resizing the DB instance class before and after each sale. Which approach fits, and how does capacity scale?

  1. Aurora Serverless v2, which measures capacity in Aurora Capacity Units and scales a writer or reader up or down in increments as small as 0.5 ACUs, without dropping existing connections during most scaling events
  2. A provisioned Aurora DB cluster with auto scaling enabled on the writer instance class, which swaps to a larger instance class within seconds of a load spike
  3. Aurora Serverless v2 configured with a single fixed ACU value equal to the peak flash-sale capacity, so no further scaling logic is needed
  4. RDS Proxy placed in front of the writer instance, which absorbs traffic spikes by pooling connections so the underlying instance class never needs to change
Next card → Shuffle