Closed shellscape closed 2 months ago
Hi @shellscape, What type of benchmarks are you hoping for? thx
I would be interested in the following benchmarks, on a matrix of 128mb, 256mb, 512mb lambdas
@Alon2303 I think @kyeotic nailed some great ones.
It'd be great, if you consider contributing to:
There's bun
, so deno
setup should be almost identical.
For you might be interested, they recently posted a new article "Benchmarking AWS Lambda Cold Starts Across JavaScript Runtimes".
In the article, Deno benchmarks using a Docker container. Node.js 20 in a docker container is ~2.4x times slower to start than the provided by AWS Node.js 20 container (using .zip).
The article is good as an attempt to persuade AWS to start providing Deno runtime for better caching and leading to better cold starts.
However, using a custom Docker image for JavaScript with AWS Lambda probably accounts to 1% of the total AWS Lambda usage for JavaScript. It’s faster to use Lambda Layers than a custom Docker container
Hi there, all that is very informative.
We're considering using AWS Lambda with the Deno runtime for a few large-scale projects, as we already have some code written in Deno, and this would help us avoid a complete rewrite in Node.js.
Performance is a key factor, important one, but we're also concerned about other potential challenges, especially at the enterprise level. Are there any issues we should be aware of, such as frequent API changes, limitations within the Deno runtime, or potential lack of support in the coming years, any shortcomings?
This repo is now archived because we are discontinuing the AWS Lambda layer for Deno. Instead we recommend you use a Docker image on AWS Lambda to run Deno. You can find a full guide in the docs.
Howdy :wave: Just wondering if you had benchmarks for Deno performance on Lambda squirreled away you could share. Trying to convince a team to give this a shot and benchmarks was the first thing they asked for.