đ¤ Judoscale News: Maximizing Your Autoscaling Performance (April 2024)
Howdy to April âď¸
Hello, Judoscalers! Time flies and weâre back to you with a new edition of the Judoscale Newsletter. With exciting news to share!
Heroku Seeks Big Fish đ
It seemed to have gone under most radars, but Heroku launched new (larger) dyno options recently! Unfortunately none of the three new dyno tiers are small â theyâre all equal to, or larger than, the Perf-L... so no win for modularity here, but these additions (along with several nods in the last year) have helped to push the idea that Heroku is looking to cater more towards enterprise customers. But hey, at least itâs clear that things are happening at Heroku and newness is coming out! Canât complain about that. Weâll keep waiting on a Perf-S or Std-4x.
Anybody switching to Perf-2XLâs?
Introducing... JudoGPT đ¤
Iâll be the first to say that there are definitely times when I donât want to talk to a human or work my way through a support-system / helpdesk to get (what feels like) a simple answer! And, even though weâve made getting in touch as easy as possible, we want to support those moments where talking with a human is just not the right move. Thus, weâre happy to announce JudoGPT! Simply put, this is a ChatGPT instance thatâs been trained on all of our docs and tooling so that it can answer your questions quickly, accurately, and with expertise. JudoGPT is available here, but also more helpfully available directly within Judoscale. Just open the Help menu!
Feel free to give it a shot and let us know what you think. Weâre always open to feedback!
Tuning Judoscale for Your App, Part 1 đ
We recently released a blog post as part of a new series all about tuning Judoscales dials and knobs for your application. We wanted to start a deep-dive into the different ways Judoscale can be tweaked to accommodate the specific needs and responses of your app; this is the first step in that journey. Itâs all about queue time.
And we talk about âqueue timeâ a lot here at Judoscale. Itâs actually the reason we exist! Being an autoscaler that operates on queue time rather than other metrics is our whole thing! But in this article we really wanted to dive into reading queue time â interpreting your applicationâs queue time metrics and making sense of it all. We believe this guide will help you tune your appâs autoscaling to reach peak performance. Please give it a read!
Maximizing Performance with Judoscale: Target Queue Time Range
An Amazon ECS Powerhouse Study
We recently sat down with the development team at TeePublic, a long-time Judoscale customer on Heroku that migrated to Amazon ECS... and took Judoscale with âem! TeePublic outgrew Heroku. Their traffic and workload was too much for 90+ Perf-L's! This is the story of how their migration to Amazon ECS, along with Judoscaleâs dynamic autoscaling, gave them the horsepower they needed to conquer Black Friday / Cyber Monday. They went from this:
To this:
And we were proud to play a role there! If youâre considering moving to Amazon ECS or even if youâre just curious what âoutgrowing Herokuâ looks like, itâs a great read.
If youâre on the other end of the spectrum and just getting a new proof-of-concept going or just giving a new tool a try, we also recently wrote a round-up guide to Herokuâs new âEcoâ dyno setup. This is the nearly-free replacement of Herokuâs actually-free dynos of yore. Eco and Basic dynos give you a ton of bang-for-your-buck and are worth exploring for all kinds of uses. Check out the guide here:
Thatâs all weâve got for you this month! Hope youâre all enjoying the spring weather out there! Weâll catch you next month with another bundle of updates and news.
Cheers!
â Jon & The Judoscale Team
P.S. Iâve mentioned our 60-second quick-start video in previous editions, but it really is worth a glance even if youâre a seasoned user. Judoscale has grown so much in the last two years that several of the features and workflows look totally different! Adamâs happy to show you around: