Exam AWS Certified Solutions Architect - Associate SAA-C02 topic 1 question 316 discussion

the problem is web server not the RDS database. My answer is C

the application run only twice a month. why I need to increase the instance size all the time, or put another server all the time? why I need to put elasticash all the time? the problem that we have only one server, it is process problem not database problem , A B D out. ANS is C.

For those who concern about "A single web server" implies ALB is useless, don't forget when you use EC2 auto scaling, when it is scaled up, which means there are more then one web server instance > minimal operational overhead. So I go with C

Answer is D. " A solution such as AWS ElastiCache and Redis can be that key component to lowering your operational overhead while providing better performance and user experience." https://cloud.netapp.com/blog/aws-cvo-blg-aws-elasticache-for-redis-how-to-use-the-aws-redis-service

A. Increase the instance size of the web server. - No, vertical scaling won't help B. Add an Application Load Balancer and an additional web server. - ALB will distribute the load evenly across multiple web servers, we can choose Weighted Round Robin Load Balancing and have the application run on AWS Beanstalk or Fargate, which will have little to no maintenance. C. Add Amazon EC2 Auto Scaling and an Application Load Balancer. - This will help, but EC2 will add more cost and maintenance effort D. Deploy an Amazon ElastiCache cluster to store frequently accessed data. - Nope

You should increase web server instances and load balance traffic through these web servers.

Timeouts are caused by servers mostly not database. since the new items are launched twice a month you will need an automated way to keep up with users requests. I'll with C

realy triky but C is elastic ....

Just Let's stick to the question. It says that "minimize" response times and timeout failures not "eliminating". I'd go for A (I agree that It still sounds a kinda lame though)

to be honest, B and C is increasing operational, even D. also the question is not mentioned its a permanent solution or not. lets stick to the requirements. the very minimal operational overhead is only vertical scaling. a few click and you're ready to go.

I would chose C for these reasons: A. "Increase the instance size of the web server" - Can be, but is vague and can not release the pressure on the DB either (caching could though..) so A - INCORRECT B. "Add an Application Load Balancer and an additional web server." - they said 400% so one server will not help and will invalidate the requirement for overhead (this will translate in this web server being used "forever" not only during those 'bad' days,) - so B - INCORRECT C. "Add Amazon EC2 Auto Scaling and an Application Load Balancer." - This will work and will add more resources when needed only, no overhead >>> CORRECT OPTION C D. "Deploy an Amazon ElastiCache cluster to store frequently accessed data." - Elasticache can help on caching and off-pressure the DB, but the use cases for Elasticache are different than "static" websites (than generate spikes only when data changes), where more a CloudFront Solution would help. So I would not chose D "Elasticache" solution unless it would be a streaming website or a realtime/media/message broker/gaming leaderboards cases - so D - INCORRECT

C is correct.

I have a question. What does it mean to "increase the instance size of the web server." Does that mean I have to deploy a new larger system with more CPU? Does AWS allow you to easily assign more CPU/memory to a VM like VMWare does? If you have to replace the server OS to "increase the instance size" then your adding operational overhead. Maybe they key word is "mitigate" in stead of solving a problem. Maybe the author of the question is used to throwing band-aids on problems to appease management with a quick pretended fix, rather than actually solving a problem.

Users encounter poor response times/timeout--improve the web tier. Best to look for autoscaling as demand is not throughout month. A does not help(vertical scaling). Plus what to do with it once demand is over. B. 1 additional server will only address 200% And it would be ideal in remaining days of no spike. C best fit as it autoscales as per demand D aim is to address web tier. Answer=C

There is 400% increase in the web traffic means, more users visit the site in first 72 hrs, so to reduce the latency and increase a response time, we simply use caching. i hope it makes sense.

Because the client wants to "minimize operational overhead", a single web server is easier to manage from an operational point of view and is a one-time action. It definitely is not the best solution both architecturally and cost-wise. A is the correct answer, IMO.

operational overhead is the key between A & C Best answer is A