Cloud Economics and SLA Questions and Answers

Answer: A

Each additional nine divides the allowance by ten, taking roughly 43 minutes down to roughly 4. At that point a single unplanned restart consumes most of the month's budget, so automated failover rather than human response becomes necessary.

Answer: B

Correct answer: 2 - 4 - 1 - 3

Ordering by the permitted downtime rather than by the percentage makes the steps clear: 99 per cent allows about 7 hours a month, 99.9 about 43 minutes, 99.95 about 22 minutes and 99.99 about 4 minutes. Each additional nine costs roughly ten times as much engineering as the one before it, which is why the target should be chosen from what the business actually needs.

Answer: C

Correct answer: Both A and R are true, and R is the correct explanation of A

Both are true and the reason is the arithmetic. Where every component must work for the request to succeed, the availabilities multiply: four services at 99.9 per cent each give \( 0.999^4 = 99.6 \) per cent, worse than any of them. That is why a service depending on many others cannot promise more than they collectively deliver, unless it can degrade gracefully or serve from cache when a dependency fails.

Answer: D

The individual figures multiply, and 0.999 cubed is about 0.997. Serial dependencies erode availability quickly, which is the arithmetic case for keeping the critical path short. Caching or making a dependency optional removes it from the multiplication entirely.

Answer: A

Correct answer: 1 and 2 only

Statements 1 and 2 are correct: a percentage measured monthly permits a different absolute outage from the same percentage measured annually, and excluded maintenance windows do not count against the figure. Statement 3 is the error the other two demonstrate - what a number means depends entirely on the window, the exclusions and how unavailability is defined, so two providers quoting 99.9 per cent may be promising quite different things.

Answer: B

Publication is common to both and tells you nothing about how they differ. The other three each change what the same headline percentage actually promises the customer. The definitions section is therefore the part of an agreement worth reading closely.

Answer: C

Raising a concurrency limit is a capacity change and will if anything increase spending. The other three are each part of the ordinary discipline of keeping a cloud bill under control. Each works by making the spending visible before the invoice arrives at month end.