DEV Community

Cover image for Let's Stop Talking About Serverless Cold Starts

Let's Stop Talking About Serverless Cold Starts

Allen Helton on September 07, 2022

I was eating dinner the other day with my family. My four year old looked at her plate and said to me "I don't want this chicken." Naturally, I as...
Collapse
 
bjorg profile image
Steve Bjorg

If it matters depends on your use-case. For synchronous interactions, such as human input, latency is a quality-of-service issue, and therefore can be critical. For asynchronous systems, it can add up and become a problem when there are a lot of hops across rarely activated functions.

I've been diving into this issue for .NET developers in a series of posts that might be of interest: Optimal Strategies for .NET on AWS Lambda. In it, I explore two strategies: "Minimize Cold Start Duration" and "Minimize Execution Cost".

Collapse
 
imflamboyant profile image
Marko Djakovic

Nice article, although there are definitely a lot of cases where often cold starts would be an issue. But that's the thing, they are most likely not that often.

I was exploring cold starts in my recent article for a specific case - AWS Lambda Cold Starts: The Case of a NestJS Mono-Lambda API

Let me know what you think about it.

Collapse
 
peteraba profile image
Peter Aba

Go is a compiled language, arguably more so than Java. It compiles to byte code instead of running on a virtual machine.