Jeff Needles (@jsneedles) 's Twitter Profile
Jeff Needles

@jsneedles

I make stuff, like @aggregationsio & @htmlcsstoimage i used to do data things for @epicgames and @whatnot

ID: 14806408

linkhttps://aggregations.io calendar_today17-05-2008 01:05:11

23,23K Tweet

4,4K Followers

1,1K Following

neuecc (@neuecc) 's Twitter Profile Photo

I've released a new compression library for C#! It provides streaming processing of LZ4 with modern .NET APIs and multi-threading. We're planning to add ZStandard with the same API. It's in preview so we'd love to get your feedback...! github.com/Cysharp/Native…

Jeff Needles (@jsneedles) 's Twitter Profile Photo

This article is not only showcasing the magic of ClickHouse but it demonstrates the right way to think about data at scale. Iterate and work on bits, don’t try to tackle everything in one go.

Dane Knecht 🦭 (@dok2001) 's Twitter Profile Photo

Triangle Man is a decent front end designer, but he knows nothing about networks. He keeps having to make this point because he has built Vercel on top of AWS which charges him for bandwidth so he has to pass that cost onto his customers or he'll go bankrupt. Cloudflare, on the

Vicent Martí (@vmg) 's Twitter Profile Photo

I don’t have strong feelings about using integers vs UUIDs for the primary key in your database. It’s a trade-off for sure. If you’re using MySQL, you should probably use integers for performance reasons, even though in practice they can be a bit annoying. VERY important gotcha

Bobby Goodlatte (@rsg) 's Twitter Profile Photo

"The Ape is gone. No utility. No community. No handshake. Just a picture of a monkey that cost me a house." This is what frustrated me at the peak of NFT mania. So many of us knew this outcome was inevitable, but very prominent VCs were bullying people online if they dared to

Aliaksandr Valialkin (@valyala) 's Twitter Profile Photo

Every read from S3 takes 100ms-200ms on average. So reading a single byte takes roughly the same time as reading 10MB of data from S3 at 100MB/s speed. This complicates designing high-performance databases with S3 storage. If the query needs reading a million of 1KB data blocks

Jeff Needles (@jsneedles) 's Twitter Profile Photo

It's great when you can cut out multiple DB round trips per request even if they're fast. In this case for @whop I added a dual layer cache for Flipper (feature flags). ~4k qps (p99 of 2ms) to ~50 qps 40%! of response times cut on some important endpoints. Gotta. Go. Fast.

It's great when you can cut out multiple DB round trips per request even if they're fast.

In this case for @whop I added a dual layer cache for Flipper (feature flags).

~4k qps (p99 of 2ms) to ~50 qps

40%! of response times cut on some important endpoints.

Gotta. Go. Fast.