AWS Lambda Silent Crash – A Platform Failure, Not an Application Bug [pdf]

If I'm reading this correctly, then AWS Support dropped the ball here but this isn't a bug in lambda. This is the documented behavior of the lambda runtime.

The document is long, and the examples seem contrived, so anyone is free to correct me but as I understand it the lambda didn't crash, after you returned 201, your lambda instance was put to sleep. You aren't guaranteed that any code will remain running after your lambda "ends". I am not sure why AWS Support was unable to communicate this OP.

If you are using Lambda with a function URL, you aren't guaranteed that anything after you return your http response remains running. I believe Lambda has some callbacks/signals you can listen to, to ensure your function properly cleans up before the Lambda is frozen, but if you want the lambda to return as fast as possible it seems you are better off having your service publish to an SQS queue instead.

This document is bizarre. The author is so confidently verbose about something they are clearly misunderstanding, and have been told as much dozens of times. It’s humbling, in a way, to think of times I’ve felt this strongly about something and to consider the possibility I could have been this wrong.

It's pretty clearly written by GPT.

Can you share some of the tell-tale signs you pick up on?

Edit: I see he's the CTO of an AI company.

He’s already added an entry about the whole incident to his resume:

https://lyons-den.com/CV/David_Lyon_CTO_CV_2025.pdf

EDIT: he has added three(!) separate mentions of the same incident to his résumé