Exam AWS Certified Solutions Architect - Associate SAA-C02 topic 1 question 312 discussion

Answer is D. if one AZ is down, left 3. IF one is used for update, left 2 AZ.

D is the answer "ensure high availability" so A and B are out since only 2 AZs "The expected request volume for the application is 100 requests per second, and each container task is able to serve at least 60 requests per second" C is wrong as you get only 60 requests per second from one task to meet 100 requests

definitely D, if one AZ is down there will still be 2 to handle the requests, one isn't enough.

D is the answer

Rolling update https://docs.aws.amazon.com/AmazonECS/latest/developerguide/deployment-type-ecs.html Rolling update Blue/Green deployment with CodeDeploy External deployment

Answer is D because you need to be able to patch, even if one AZ is down. You have to have two container "tasks" up in an AZ to process the "100 requests per second". Thus you still have to have one AZ up with two containers up at all times, even with one AZ offline, and containers in the other being patched. This still doesn't leave you with HA if the patching goes bad and you lose the patched containers because of it. Good HA costs more money. Based on the criteria, it requires three zones at a minimum, and really four if you wanted to maintain HA at all times if you lose an AZ and still want to patch. In the real world, if I lost an AZ all patching/changes would STOP until until the lost AZ was back up and stable.

1000000000%D

total 6 -2(1 AZ down) -1(update) =3 3/6=50%

Answer: D 3 AZ, and 2 Tasks in each 1 AZ went down, so we have 2 AZ with 4 tasks As rolling update is there with minimum 50% capacity - 2 tasks will go down for replacement. Leaving 2 more tasks -giving 120 requests/second. Hence, D is the answer.

" The company set up Amazon ECS with a rolling update deployment type with the minimum healthy percent parameter set to 50% and the maximum percent set to 100%." there are 2 instances, if 1 is down, auto scalling will not work as the minimal health is set to 50%

I am thinking from 200 onwards many answers are correct, and we are over-thinking. See they asked for high availability so workload can run at 50% (minimum healthy) of course degraded but it will run. If they ask for Fault Tolerance then system should run at 100% capacity even after failure. Please search for HA and FT on aws. If one AZ is down they have configured minimum health as 50% and ECS with rolling update, hence one task in one AZ is still there. "HA"

A or D?

The 50% minimum healthy limit is critical here. IT means worst case we may have only 1 task running out of 2 tasks in order for that AZ to be up

typical 3-zone architecture. There is always one in maintenance mode for updating, two others to provide HA. Once the update is done, rotate out another zone for updating until all three zones done.

I have not found the right answer. But this question doesn't say most cost-effective or minimum. So I will choose D.

The answer is D 100%

Minimum set to 50 and Maximum set to 100. If one AZ is down we still have 2 AZ's and remember we need 2 AZ's with a worst case scenario of 50% on each AZ......Ans: D