[pavelrub]

Specifying the max number of concurrent lambda execution has been possible for years - it's called reserved concurrency.

This does nothing to solve the issue however. There are many use cases where you can't delay handling requests.