I have started seeing Memory Size: 128 MB Max Memory Used: 128 MB messages in my CloudWatch logs. I would expect that when the limit is reached the lambda should exit with an error. However, this is not happening. Aside from the above message, there is no indication that the lambda failed.
Does anybody know more about what is happening under the covers in these cases? Is the lambda still running and maybe AWS is doing some magic to manage the memory so it doesn't fail, but maybe the lambda just takes longer to run?
Or is it failing without any log or notification? Or something else entirely?