Exam AWS Certified Solutions Architect - Professional SAP-C02 topic 1 question 120 discussion

API Gateway has a limit of 10k requests per second, per account, per region https://docs.aws.amazon.com/apigateway/latest/developerguide/limits.html

The correct answer is C. The company reached its API Gateway account limit for calls per second. This is because Amazon API Gateway has a default account-level limit of 10,000 requests per second (RPS) and a default per-method limit of 5,000 RPS. If the company's premium tier customers are making more than 10,000 requests per second in total across all API methods and regions, they would be receiving the error message of 429 Too Many Requests. This indicates that the API Gateway account is reaching its capacity limit, and the Lambda function is not being invoked because API Gateway is blocking the requests before they reach the Lambda function. The other choices are not correct because the Lambda function's concurrency limit and region limit for concurrency would not affect the API Gateway's request rate limit, and the API Gateway's default per-method limit is 5,000 RPS which is less than the premium tier's 3,000 calls per second.

The most likely cause of the 429 Too Many Requests error messages, despite the Lambda function not being invoked, is D. The company reached its API Gateway default per-method limit for calls per second. Here's a breakdown of why: API Gateway Limits: API Gateway imposes rate limits on API methods to prevent abuse and ensure fair resource allocation. These limits can be configured at the account level or the individual method level. Default Limits: If not explicitly configured, API Gateway applies default limits to methods. These default limits may be insufficient for high-traffic scenarios, especially during peak usage hours. Lambda Function Invocations: The Lambda function is not invoked because the request is being throttled at the API Gateway level before it reaches the Lambda function.

C. The company reached its API Gateway account limit for calls per second.

I'm going to argue the problem is that the source of the errors is Lambda reaching it's regional concurency limit. By default, Lambda has a regional limit of 1000 concurrent invocations. The premium tier allows 3000 requests/s. Depending on the number of premium customers and the average duration of a call it may be that the concurency limit is reached or not. We don't know for sure, but it is certainly plausible. If Lambda hits the concurrency limit, it also returns 429. So how do we know where the error is coming from? The key is in what exactly fails: "multiple API methods during peak usage hours". Notice it is not ALL API calls. If the gateway was rate limiting then we would see blocks of random requests being denied until the rate bucket empties to allow new ones. But if the Lambda concurrency is hit then it means no new execution environments of Lambda are created, but the ones that exist keep processing. So the behaviour is that some functions will continue to operate, while others will start thorttling with 429. This behaviour better matches with "multiple API methods" failing.

C correct. This is Gateway API response

C is correct.

429 is API Gateway API throttle default limit.

C of course

C 429 error indicates that API calls per second was exceeded ... it's not a Lambda issue

Company reached its limit

C is the answer

C is correct answer