Re:AWS Lambda Silent Crash – A Platform Failure, Not an Appl...
by Everdred2dx on Monday July 14, 2025 @11:33PM
(#44567619)
They mentioned “rebuilding” in EC2. Probably just moved that over to Azure and no longer are bothering with serverless :)
Re:AWS Lambda Silent Crash – A Platform Failure, Not an Appl...
by pryelluw on Tuesday July 15, 2025 @12:03AM
(#44567724)
I mean, they should just have created another queue, that triggers another lambda that then continues the process after the http response is sent.
However, it is still funny when you think that in regular servers you still finish when you respond to the request.
My dog keeps getting angry at the door. It won’t open automatically.
Re:AWS Lambda Silent Crash – A Platform Failure, Not an Appl...
by arpinum on Tuesday July 15, 2025 @05:06AM
(#44569315)
Azure functions don't operate in the same way, they are more like AppRunner's architecture. They then layer on the Lambda billing model.
It was typical Azure strategy of rushing out a product to show they have all the same features as AWS. It may have improved since I last touched it in 2018.
Re:AWS Lambda Silent Crash – A Platform Failure, Not an Appl...
by whatthemick on Tuesday July 15, 2025 @09:55AM
(#44571211)
As of 2025 it's still very much the same.
Re:AWS Lambda Silent Crash – A Platform Failure, Not an Appl...
by jiggawatts on Tuesday July 15, 2025 @05:48AM
(#44569545)
> That’s like expecting a local function to continue executing after a `return` statement.
Well... they can![1]
It's definitely a thing in traditional always-running web server processes, which is all people know these days.
Many developers have never written a code for a short-lived process outside of maybe school exercises. Things like CLI tools, background jobs, etc...
[1] I get "Later!" printed out... most of the time: https://dotnetfiddle.net/ng93d4
Re:AWS Lambda Silent Crash – A Platform Failure, Not an Appl...
by williamdclt on Tuesday July 15, 2025 @10:31AM
(#44571542)
Also, `defer` in go or `finally` in JS/Python
But one thing bothers me…wouldn’t the author have encountered the exact same behavior in Azure? I guess I really will have to read this paper to find out.