Doubling compute doesn't always mean doubling capability. Talking about scalability, sharding, and handling unexpected requests in the latest yt.

@chrismunns I like to go wide, we’ve got a few dozen 1gb 1vcpu containers running our web traffic. Lots of nodes >>> big node

Top-tier content, as usual Real Software Engineering by a real Software Engineer

thanks for covering up my weekend learning topic ben

Have y'all talked to AWS? I feel like you're touching on a lot of the design principles they already had to deal with for example in Redshift. Matt Abrams and Rahul P. And perhaps Ippokratis P. Would be worth talking to especially around workload management and shared task models

thanks for the video

nice video. learnt alot.