Maxime David

Feeling truly honoured to be featured on the Rust youtube channel 🥹! Here is the replay of a talk I gave last month about Rust and Serverless at @rustlinz in case you want to 👀

Peek behind the curtain of AWS Lambda! Join us with @_maxday (Maxime David) from the Lambda runtime team to discover how AWS manages millions of serverless functions while keeping them secure and up-to-date. Listen now in your podcast apps 👇🧵 #AWS #Serverless

@astuyve Happpppy birthday 🎂! 🐶 ❤️

it's my birthday, say something nice about datadog

Thanks for the mention 🤓😇

It was a great conversation between @edjgeek and @_maxday. Thanks for mentioning article "Improving Lambda performance with the help of AWS Lambda Profiler Extension for #Java" (dev.to/aws-heroes/aws…). #AWS #Serverless #performance #profiler twitch.tv/videos/2439808…

@astuyve 😂😂😂

My baby woke up screaming so I started telling her about NAT-punching for p2p applications and suddenly not only is baby asleep, but so is my wife AND father in law?

@mattzcarey @siiiaaaaaaa @rustlinz I’m curious now, can you tell me more about your project?

@siiiaaaaaaa @_maxday @rustlinz you’d be surprised. Actually a pretty good use case for high scale lambda

Feeling truly honoured to be featured on the Rust youtube channel 🥹! Here is the replay of a talk I gave last month about Rust and Serverless at @rustlinz in case you want to 👀

@mattzcarey Hey! Lambda timeout is 15min max. Not sure that would be the perfect use case for server sent events? What are you trying to build? There is a nice integration with API gateway and websockets if that would work for you

mad that lambda still doesn’t support returning server side events to a client. you need a crazy websocket setup or rely on the terrible response streaming. @_maxday can you fix it?

@_maxday Does it count if I run the Lambda runtime interface emulator on the box?

it is with great sadness that I must inform you all that I've booted an EC2 instance into my "serverless" AWS account.

@astuyve Awwwww congratulations to both of you 🥰🥰

Our baby girl arrived early this morning! Everyone is recovering well. Glad I sorted out the car seat yesterday!

@manofyear93 @loige @taktakpeops aww thanks for the mention @loige

@loige @taktakpeops @_maxday Brill! Thanks for the replies!

Why you should consider Rust for your Lambdas loige.co/why-you-should…

@astuyve @loige @MarioVerbelen Yeah we looked at tinygo, unfortunaly it was not a great fit for our use case (mainly because some packages are not yet supported, see tinygo.org/docs/reference…) That said, I think it clearly deserved a spot in lambda-perf! I guess I found what to do this week-end :) thanks @loige

@loige @MarioVerbelen @_maxday Max has looked at tinygo extensively, so I don't want to speak for him on that. I took up the other side of those questions and explored a ton of fun side quests like tuning GC or disabling it entirely for Lambda

Okay you probably don't need Rust in Lambda anymore, sorry for the confusion everyone. Just javascript and LLRT:

@tr_twtr @sampullara @astuyve Sure, everything is fully open-sourced, feel free to check the corresponding runtime: github.com/maxday/lambda-… Let me know if you have any questions!

@sampullara @astuyve @_maxday True, maybe @_maxday could make it more clear there.

Let's go 🚀🚀 New llrt runtime has some pretty nice cold start performance. Will we see this as a managed runtime in Lambda??

@_maxday @thdxr Pretty impressive to see on par or slightly better performance than Go (of course with a much more limited feature set) Thanks for updating the benchmarks 😊

My pleasure! 😇

@thdxr it's live :)

@thdxr github.com/maxday/lambda-…, will be live on the benchmark tomorrow 9AM EST

looking through aws llrt[1] and i'm impressed this is the exact type of thing i always complain that aws is not doing but should be some details and initial thoughts: tldr it's another js runtime but with a focus on starting up fast for short lived work - perfect for serverless functions this should remind you of cloudflare's workerd it's extremely small < 3mb so you can drop it into a lambda function today as a custom runtime and it'll work well built into these 3mb is also a bunch of the most popular aws-sdk clients implemented natively so you won't have to bundle those unlike deno or bun, which are general purpose runtimes, llrt is focusing on serverless environments this lets it drop things like JIT which allows it to be smaller as with any js runtime effort not all the nodejs apis are available yet but coverage looks pretty good already claiming 10x faster to start, 2x faster to run this is all fantastic - betting on a specific (but potentially huge) workload and optimizing it all the way down i don't think this will be used directly too but it provides frameworks like sst or middlemen like triangle co an efficient setup to target on behalf of end users

@thdxr Ahah :) it should be live tomorrow on lambda-perf ❤️