3

For a stage belonging to an API in AWS API Gateway I have the option to limit Default Method Throttling. Does this limit the total number of requests per second, or the number of requests from a particular client per second?

Jansky
  • 1,455
  • 1
  • 17
  • 33

1 Answers1

4

Default Method Throttling (like Account Level Throttling) is the total number of requests per second across everyone hitting your API.

Client-level limits are enforced with Usage Plans, based on api-keys.

For more detailed information about API Gateway throttling checkout:

thomasmichaelwallace
  • 7,926
  • 1
  • 27
  • 33
  • I found the formulation in the aws-docu very misleading: "Throttling and quota **limits apply to requests for individual API keys** that are aggregated across all API stages within a usage plan". In reality the limits are **shared across all API-keys within a usage plan**. Double confirmed that with AWS support. – Skip Mar 31 '23 at 10:04
  • @Skip that isn't correct. The limits aren't shared by all api keys within a usage plan. They are shared across apis per usage plan + api key combination. – theawless May 17 '23 at 00:23
  • @theawless have I understood you correctly: when creating a usage plan, and assigning an API X to usage-plan, and creating 2 api-keys K1, K2 THEN when **calling API X with K1** would give me independent throttling counter, of when **calling API X with K2** – Skip May 23 '23 at 07:10
  • Yes that's correct, keys always get different counters. With your example if you create API Y and attach it to the usage plan as well, then "are aggregated across all API stages within a usage plan" would mean calling API X with K1 and API Y with K1 would get a shared counter, but API X with K1 and API X with K2 will get different counters. – theawless May 24 '23 at 23:59