Exam AWS Certified Solutions Architect - Associate SAA-C02 topic 1 question 433 discussion

Non-critical overnight workload is SPOT. Since the question says overnight, we will assume its non critical specially its just inventory, they can redo it again, so it must be A or D only. It needs SQS messaging for the worker node, so its D then.

D https://aws.amazon.com/vi/blogs/compute/running-cost-effective-queue-workers-with-amazon-sqs-and-amazon-ec2-spot-instances/

C is correct and the only option which ensure metadata get process by worker node. D is cheap but subjective to spot availability/interruption for metadata processing. https://aws.amazon.com/vi/blogs/compute/running-cost-effective-queue-workers-with-amazon-sqs-and-amazon-ec2-spot-instances/

Spot Instance for cost saving

All explanations are misleading. The strong doubts are between C (on demand) instance and D (spot instance in auto scaling). Perhaps D wins, despite the phrase "must guarantee that worker nodes process each picture" because all images to process are stored in S3. So we could assume that it's a batch processing that could use spot instances (MOST cost effective) and, in case of maximum SQS retention (14 days) it' enough to re-process the missing images still in S3 when the spot instance are available again. So something that read images from S3, write them in SQS and process by Spot Instances in Auto scaling. If something go wrong it's possible to read again missing images from S3. In any case, I expect that on the exam such a question will be written much better

READ THIS ARTICLE : https://aws.amazon.com/vi/blogs/compute/running-cost-effective-queue-workers-with-amazon-sqs-and-amazon-ec2-spot-instances/

Answer is D

guuuys it's D

A solutions architect must guarantee that worker nodes process each picture ---> we use queue to keep messages so A & B is out met in the MOST cost-effective manner possible --------> Spot instances cheaper than on-demand

If the load is critical, there is a chance of all the spot instances in a group being terminated because of the demand. The ideal would be to have a mix of on-demand/Reserved instances and spot instances in an auto scaling group. But since that's not an option, D is the next best option which checks most of the boxes required by the question, according to the link that someone posted in the comments. https://aws.amazon.com/vi/blogs/compute/running-cost-effective-queue-workers-with-amazon-sqs-and-amazon-ec2-spot-instances/

The keyword is "MOST cost-effective manner" So you use SQS to make sure the message will never lose, you can use spot instance to cut cost. So D

I picked B and I stick by it. It's the most cost-effective solution. Directly to autoscaled EC2s ensures the image is always processed and it's using reserved instances which also saves a lot of money. Don't over complicate it!!

Answer is D "a custom Amazon CloudWatch metric for the current number of messages in the queue. " it is a right way to scale

DDDDDDDDDD

D. A-spot instances can be lost moment spot price crossed our fixed price--no 'guarantee'(even if ASG spins new instance,that particulat job is lost) B-fits,but RIs mean we are still paying irrespective of what ASG does C-fits (less scalable/more costly as only doing protection, not auto scaling using policy) D- best option, SQS has the concept where consumer(workers)must manually delete the message post processing. If Spot instance gets lost,job is still intact.

C - C makes sure that each and every image is processed. Not using Spot ensures that none are missed.

D - cause of https://aws.amazon.com/blogs/compute/running-cost-effective-queue-workers-with-amazon-sqs-and-amazon-ec2-spot-instances/ Applications can minimize the impact of a Spot Instance interruption. To do so, an application catches the two-minute interruption notification (available in the instance’s metadata), and instructs itself to stop fetching jobs from the queue. If there’s an image still being processed when the two minutes expire and the instance is terminated, the application does not delete the message from the queue after finishing the process. Instead, the message simply becomes visible again for another instance to pick up and process after the Amazon SQS visibility timeout expires. Alternatively, you can release any ongoing job back to the queue upon receiving a Spot Instance interruption notification by setting the visibility timeout of the specific message to 0. This timeout potentially decreases the total time it takes to process the message.