Exam AWS Certified Solutions Architect - Associate SAA-C02 topic 1 question 375 discussion

In this case, looking at various options. A. This will work but there is no reliability as message can be lost. B. Lambda is managed service that adjust itself with the load, but you cant use autoscaling. C. Not sure about App Mesh, which is a service that provides application level networking. D in this case works, without a doubt. The requests from the API are sent to the models Amazon Simple Queue Service (Amazon SQS) queue. - Valid and will work Models are deployed as Amazon Elastic Container Service (Amazon ECS) services reading from the queue - Make sense AWS Auto Scaling is enabled on Amazon ECS for both the cluster and copies of the service based on the queue size - This makes absolute sense.

D is the answer, Lambda can't process due to the model size(1GB), directory memory allocation is 512MB.

A is wrong, question mentioned "asynchronous", so we need a queue B is wrong NOT because of the 512MB size limit of Lambda (the limit increased to 10GB in late 2022), but because we CAN'T directly increase vCPUs of Lambda, we can only increase memory of Lambda (as memory is increased, the number of vCPU also go up - 1 vCPU per 1769MB) C is wrong, App Mesh is not for managing auto scales So the only one left is D https://docs.aws.amazon.com/lambda/latest/dg/configuration-function-common.html

D is the answer

The solution that meets these requirements is option B. The requests from the API are sent to the models Amazon Simple Queue Service (Amazon SQS) queue. Models are deployed as AWS Lambda functions triggered by SQS events AWS Auto Scaling is enabled on Lambda to increase the number of vCPUs based on the SQS queue size. By using Amazon SQS, the company can manage the requests to the models in an asynchronous manner, which can help to handle the irregular usage patterns. The models are deployed as AWS Lambda functions, which can be triggered by the events in the SQS queue, and automatically scale based on the size of the queue with the help of AWS Auto Scaling. This way, the company can handle the irregular traffic patterns and ensure that the models are available to users even during high traffic spikes.

B is not right, since "Models are deployed as AWS Lambda functions triggered by SQS events" can be true. SQS message only can be polled, not push.

lambda scales automatically, has to be D

D is correct. No autoscaling for Lambda

DDDDDDDDDDDDDD

B is the answer. SQS helps in de-coupling Lambda ensures there are no EC2 or ECS running without working on models for weeks The last sentence in answer B about auto-scaling https://docs.aws.amazon.com/autoscaling/application/userguide/services-that-can-integrate-lambda.html

ALB inovke Lambda is synchronous. But it mentioned that "Users access the models through an asynchronous API" So A is not right

No point of having models loaded to ECS as some of the models are not used for weeks. Lambada Memory Limit - https://aws.amazon.com/about-aws/whats-new/2020/12/aws-lambda-supports-10gb-memory-6-vcpu-cores-lambda-functions/#:~:text=relevant%20marketing%20content.-,AWS%20Lambda%20now%20supports%20up%20to%2010%20GB%20of%20memory,vCPU%20cores%20for%20Lambda%20Functions&text=AWS%20Lambda%20customers%20can%20now,previous%20limit%20of%203%2C008%20MB. Lambda Scaling - https://docs.aws.amazon.com/lambda/latest/dg/invocation-scaling.html So it must be either A or B.

What is wrong with C?

Ans: D I will go with option D, Models are built on microservice, workloads are unpredictable.

Question is not clear to me. If all model data is 1Gb then it is A, if each one is 1GB then answer is D.

Those who think Lambda cannot be scaled, please look at this link -https://docs.aws.amazon.com/lambda/latest/dg/invocation-scaling.html. Also, the question says that 'Some models could be unused for days or weeks'. I wouldn't want to pay for those ECS containers when models on those containers are not used at all even though Contrainers in that ECS environemnts are running with minimum instances. I'd just go with Lambda which only triggers when there is a load and RAM requirements are also met. I'd go with B but happy to be corrected with logic and explanation with links.

A is out for sure B is also out since you don't scale Lambda I'm taking C over D https://aws.amazon.com/app-mesh/?aws-app-mesh-blogs.sort-by=item.additionalFields.createdDate&aws-app-mesh-blogs.sort-order=desc&whats-new-cards.sort-by=item.additionalFields.postDateTime&whats-new-cards.sort-order=desc: "You can use App Mesh with AWS Fargate, Amazon EC2, Amazon ECS, Amazon EKS, and Kubernetes running on AWS, to better run your application at scale."