The city's nonproduction account mirrors production sizes and runs all night and weekend. Which change best reduces nonprod compute cost while keeping weekday testing?
Select an answer to reveal the explanation.
Short Explanation
QA doesn't need parade-day horsepower at 3 a.m. Shrink the boxes versus prod and put them on a nights-and-weekends off switch—your budget will thank you Monday morning.
Full Explanation
Using smaller instance classes than production and scheduling instances to stop nights and weekends is the correct nonproduction cost reduction while keeping weekday testing, because QA and lower environments seldom need production horsepower at 3 a.m. Municipal nonprod accounts that mirror production sizes around the clock burn budget on idle capacity; Instance Scheduler, EventBridge, or ASG schedules reclaim those hours. Requiring multi-AZ active-active for every developer desktop VPC fails by imposing production-grade HA cost on nonprod desktops that do not need that topology. Purchasing three-year All Upfront Reserved Instances for idle QA boxes only fails because long RI commitments on fleets that should be stopped most hours locks spend into the wrong shape of waste. Moving nonprod to a second commercial cloud without tagging fails conceptually: changing providers without tagging and scheduling discipline does not fix idle runtime and harms cost attribution. Exam caveat: shrink and schedule nonprod versus prod; do not buy long RIs for idle QA or escape accountability with untagged multi-cloud moves. Operational check: rightsize nonprod instance classes; enforce stop schedules outside business hours with exceptions tagged; verify weekday test windows; report weekend running instances; and compare nonprod spend month over month for the city’s accounts.