Exam AWS Certified Solutions Architect - Associate SAA-C03 topic 1 question 516 discussion

In the context of the given scenario, where the company wants low latency and consistent performance for their API during peak usage times, it would be more suitable to use provisioned concurrency. By allocating a specific number of concurrent executions, the company can ensure that there are enough function instances available to handle the expected load and minimize the impact of cold starts. This will result in lower latency and improved performance for the API.

Provisioned - minimizing cold starts and providing low latency.

Concurrency of lambda function, provisioned vs reserved: Reserved Concurrency: 1. Desired instances can start without interference from other functions 2. Guarantees available execution capacity for critical functions Provisioned Concurrency: 1. Also guarantees consistent performance 2. [!!] Allows you to pre-warm a specific number of function instances to reduce cold start latency

https://docs.aws.amazon.com/lambda/latest/dg/lambda-concurrency.html#reserved-and-provisioned Consistency decreases if you exceed your provisioned instance. Lets say you have 1000 (default) provisioned instances and the load is 1500. The new 500 will have to wait until the first 1000 concurrent calls finish. This is solved by increasing the provisioned concurrency to 1500.

So I have my doubts here. The question also states ;"The company needs to provide a compute host for the API." Imho this implies to have some sort of physical host which has to be provided by the customer. Translating this further to aws this would mean an EC2 instance. And then when I would go for ECS in stead of EKS. Please share your opinion.

This option provides the least operational overhead: API Gateway handles the API requests and integration with Lambda Lambda automatically scales compute without managing servers Provisioned concurrency ensures consistent low latency by keeping functions initialized No need to manage containers or orchestration platforms as with ECS/EKS

The company requires the API to respond consistently with low latency to ensure customer satisfaction especially during high peak periods, there is no mention of cost efficient. Hence provisioned concurrency is the best option. Provisioned concurrency is the number of pre-initialized execution environments you want to allocate to your function. These execution environments are prepared to respond immediately to incoming function requests. Configuring provisioned concurrency incurs charges to your AWS account. https://docs.aws.amazon.com/lambda/latest/dg/provisioned-concurrency.html#:~:text=for%20a%20function.-,Provisioned%20concurrency,-%E2%80%93%20Provisioned%20concurrency%20is

AWS Lambda provides a highly scalable and distributed infrastructure that automatically manages the underlying compute resources. It automatically scales your API based on the incoming request load, allowing it to respond consistently with low latency, even during peak times. AWS Lambda takes care of infrastructure provisioning, scaling, and resource management, allowing you to focus on writing the code for your API logic.