An organization runs a Presto engine in watsonx.data that is heavily used during business hours and idle overnight. They want to reduce compute cost without losing the engine's configuration, its catalog associations, or the data it serves. What action in Infrastructure Manager fits that requirement?

The correct action is pausing the engine outside business hours and resuming it when the workload returns. Infrastructure Manager exposes pause and resume for engines so compute can be stopped when it is not needed. The engine object, its size, and its catalog associations all survive the pause, so resuming restores the exact serving configuration without any reconfiguration step. Deleting the engine each evening and recreating it in the morning from the same settings is incorrect because deleting an engine discards its associations. Recreating it each morning means re-associating catalogs every time, which is manual work that will eventually be done wrong, and it gives you no rollback if the recreation is misconfigured. Moving the underlying bucket to an archive storage tier every night is incorrect because it confuses compute cost with storage cost, and it targets the wrong one. The overnight spend here is the running engine, not the object storage. Worse, moving actively queried data to an archive tier introduces retrieval latency and cost the moment business hours resume. Reducing the engine to a single worker node permanently and accepting slower daytime queries is incorrect because it solves the cost problem by permanently degrading the workload it exists to serve. Right-sizing an engine is legitimate, but sizing for the idle period and living with it during peak is not right-sizing. For the exam, associate pause and resume with cost control on a predictable duty cycle, and scaling with sustained changes in workload.