iiinova — where your digital self keeps evolving

Adam McCrea @adamlogic.com · 13/05/2026

Render made it SO EASY to connect logs and metrics to a third-party observability tool. Railway, as far as I can tell, doesn't support either. 😞 What am I missing?

000

Adam McCrea @adamlogic.com · 04/05/2026

Render might not be a great option for high-traffic apps using a separate hosted DB. (This is from our own prod app)

240

Adam McCrea @adamlogic.com · 06/04/2026

We REALLY want to know what's going on at Heroku ...so we wrote them a letter 🤷‍♂️

140

Adam McCrea @adamlogic.com · 23/03/2026

Heroku is doubling down on the "sustaining engineering" language and it really bothers me. For most devs, "sustaining engineering" == "maintenance mode" If they don't want customers to think they're in maintenance mode, they need clearer language.

150

Adam McCrea @adamlogic.com · 18/02/2026

Thoughts on env vars as I spend some time with Fly... I actually like that Fly separates env vars and secrets as distinct concepts. Confused me at first, but it makes sense. Secrets get extra protection, env vars are for config. Good semantic separation.

110

Adam McCrea @adamlogic.com · 18/02/2026

I love how Northflank's pricing page lets you choose the timeframe for the metered pricing. All metered pricing should have this.

000

Adam McCrea @adamlogic.com · 16/02/2026

I just updated our PaaS Pricing Comparison Calculator: - Added Amazon ECS Fargate - Updated pricing for Railway - Factored CPU time-slicing for Fly shared machines Heroku still stands out as the most expensive, but the others are very competitive!

120

Adam McCrea @adamlogic.com · 03/11/2025

On @searls.bsky.social latest podcast he mentioned "Heroku" and "price", so I can't help but respond. I made the post public since I think it's a really useful pattern. tldr: Got a high-memory job that pushes you to perf dynos? Isolate it and scale to zero.

121

Adam McCrea @adamlogic.com · 15/09/2025

Nearly 10 years (!!) of autoscaling, some patterns never change: noisy neighbors, too many queues, scaling workers to zero (and saving 💰), and the chaos of downscaling too far. Stability is boring, repeatable choices. Here's what we've learned 👉 judoscale.com/blog/autosc...

051

Adam McCrea @adamlogic.com · 14/07/2025

If I was responsible enough to do any sort of ROI calculation, this never would have happened. Instead, we went all-in on RailsConf with 15 homemade videos and a custom local-only app to run our booth. I'd do it all over again.

160

Adam McCrea @adamlogic.com · 10/07/2025

As requested by absolutely no one, a supercut of every kazoo clip we made for #railsconf2025

2353

Adam McCrea @adamlogic.com · 09/07/2025

Really hope we don’t get kicked out of the last #railsconf for bringing kazoos 🎶

020

Adam McCrea @adamlogic.com · 08/07/2025

Visit our booth at #railsconf to get your kazoo! Watching us make fools of ourselves is just a bonus.

020

Adam McCrea @adamlogic.com · 04/07/2025

If it’s not at least a little bit cringe, we didn’t go far enough. #railsconf #kazoos #marketing

010

Adam McCrea @adamlogic.com · 03/07/2025

Marketing technique #77: Humiliate yourself for attention. #railsconf #kazoos

020

Adam McCrea @adamlogic.com · 02/07/2025

Y’all this is only the beginning. Buckle up! #railsconf #kazoos

030

Adam McCrea @adamlogic.com · 01/07/2025

Most autoscalers are reactive—they respond to traffic after it spikes. That’s usually fine… until it’s not. If your app gets hit with big surges of traffic all at once, even the fastest queue-time autoscaler might not spin up capacity in time to avoid slowdowns or timeouts.

110

Adam McCrea @adamlogic.com · 01/07/2025

One week until #railsconf! The whole Judoscale team will be there with... kazoos?

080

Adam McCrea @adamlogic.com · 29/04/2025

Python task queue pro-tip: Autoscale based on queue latency, not CPU! Your task queue can back up without CPU spikes, leaving you in the dark. Check out Jeff's full Celery & RQ comparison here: judoscale.com/blog/choose...

010

Adam McCrea @adamlogic.com · 29/04/2025

Choosing a Python task queue? Jeff Morhous compared Celery vs RQ: 🧠 Celery: Feature-rich but complex 🚀 RQ: Simple & easy to deploy Jeff's advice: Most apps do fine with RQ until they need more horsepower. Then consider making the switch to Celery.

110

Adam McCrea @adamlogic.com · 22/04/2025

This behavior is confusing at first, but it's actually super cool once you wrangle it. And with autoscaling in place, we don't really need to worry about it. New machines are being created with fresh burst balances when needed, spreading the load allowing balances to rebuild.

010

Adam McCrea @adamlogic.com · 22/04/2025

But why do they perform so well at first, only to fall apart after a bit? It's the bursting! Shared machines have a "burst balance" where they can use 100% CPU. Once the balance is depleted, CPU is throttled to 1/16.

110

Adam McCrea @adamlogic.com · 22/04/2025

I finally answered my questions about shared CPUs on Fly.io: - How are they so cheap? - Why do I need so many of them? - Why does perf tank after 10 minutes? Turns out it's well-documented: "shared" machines only get 1/16 of each CPU! fly.io/docs/machin...

110

Adam McCrea @adamlogic.com · 21/04/2025

⚠️ It's a mistake to ignore these configs. The defaults are NOT what you want. EXAMPLE: A 4-process, single-threaded web server should use a hard limit of 4 since that's the max concurrent requests. A soft limit of 1-2 would help route requests to less busy machines.

100

Adam McCrea @adamlogic.com · 21/04/2025

Last week I dug into HTTP routing behavior on Fly.io, and it's so cool! Unlike random routing on platforms like Heroku, Fly can intelligently route requests to machines based on a load. Here are the configs you need to know...

251

Adam McCrea @adamlogic.com · 17/04/2025

If you have more than one worker process sitting idle while the job queue is empty, you're throwing away money. And for queues with an SLO > 30 seconds, that number drops to zero.

053

Adam McCrea @adamlogic.com · 16/04/2025

"Does Rails 8 truly make platforms a thing of the past? And can small to medium-sized teams really save money by self-hosting?" Preech it, Jeff! From the Honeybadger blog: www.honeybadger.io/blog/rails-...

240

Adam McCrea @adamlogic.com · 16/04/2025

Feeling very thankful today as Carlos Antonio and I celebrate his one year anniversary at Judoscale. 🥳 Carlos helps me level-up as a developer and feel more relaxed and confident as a founder. And with him Judoscale is a better product than ever. Thanks, Carlos! 🥂

030

Adam McCrea @adamlogic.com · 09/04/2025

Related: Jon Sully's similar guide as the final part of his 3-part propshaft series judoscale.com/blog/propsh...

010

Adam McCrea @adamlogic.com · 09/04/2025

I love this deep dive into importmap and propshaft from Radan Skorić We use both at Judoscale, and even though each of these are simple in isolation, it's crucial to understand how they work together. radanskoric.com/articles/ra...

162

Adam McCrea @adamlogic.com · 08/04/2025

In this example, 5_sec_worker is always running. The others autoscale down to zero, and scale up based on their SLO. Each queue is isolated, meaning one queue can't starve another. Simple, safe, efficient.

010

Adam McCrea @adamlogic.com · 08/04/2025

"One queue per worker" is the way. Sidekiq, Solid Queue, doesn't matter. In other words, each queue should have a dedicated worker process. But isn't that wasteful?? Nope. You autoscale them down to zero when they’re idle. They’ll start back up when jobs are enqueued.

150

Adam McCrea @adamlogic.com · 07/04/2025

I've really been into Slack "canvases" lately. Here's how we use them to track & plan our content publishing calendar as a small team. It's just a simple doc with a permanent tab in our #marketing channel. No need for a full calendar in Notion or anything like that. KISS

020

Adam McCrea @adamlogic.com · 04/04/2025

Our sleuthing continues... it looks like something changed on July 16, 2021... but no associated changelog entry, so no idea what that change was.

000

Adam McCrea @adamlogic.com · 19/03/2025

Exciting to see autoscaling in action for our latest integration with Fly.io! (because CPU is a poor metric for queue capacity)

Screenshot of autoscaling Fly.io worker machines based on job queue latency

0100

Adam McCrea @adamlogic.com · 18/03/2025

Help me fix our pricing page! I've had multiple people tell me our "freemium" plan is not clear. Like, they didn't even realize we HAD a free plan. I can see a couple issues, but what else? How can I make it clear??

331

Adam McCrea @adamlogic.com · 11/03/2025

I'm so excited to share the tool I've been working on for the past month! Do you know how much it would cost to host YOUR app across Heroku, Render, Fly, and Railway? Now you can compare costs side-by-side. judoscale.com/tools/paas-...

74310

Adam McCrea @adamlogic.com · 03/03/2025

📣 New Judoscale feature We've expanded your autoscaling and queue time history from the past day to the past week!

041

Adam McCrea @adamlogic.com · 28/02/2025

We have a winner in our Sin City Ruby giveaway! Congrats Matt, and thanks @jwright.bsky.social for playing.

131

Adam McCrea @adamlogic.com · 10/02/2025

Come hang with me at Sin City Ruby in April! We're giving away a free ticket INCLUDING HOTEL!! You just need to get yourself there. 😁 @jasonswett.bsky.social puts on an amazing show (so I've heard), and Judoscale is excited to be a sponsor. Details here: judoscale.com/blog/sin-cit...

0147

Adam McCrea @adamlogic.com · 15/01/2025

It looks like Redis Cloud doesn't even publish their pricing anymore. Am I missing something? If not, that's very disappointing.

110

Adam McCrea @adamlogic.com · 14/01/2025

And the theme song has already been written.

010

Adam McCrea @adamlogic.com · 10/01/2025

Working on some new diagrams for one of our guides. Visuals like this are so important to effectively convey techincal concepts. Sure, they take time (usually more time thinking than actual drawing), but they're SO WORTH IT! 💅

140

Adam McCrea @adamlogic.com · 03/01/2025

Getting R14 errors with Rails on Heroku? Here's a quick win that takes 2 minutes to implement: Add the Jemalloc buildpack to your app and set JEMALLOC_ENABLED=true. That's it! We've had huge memory savings with zero downside. More in Jon Sully's article: judoscale.com/blog/rails-...

1173

Adam McCrea @adamlogic.com · 02/01/2025

Ever tried to "save photo" from Giphy and got WEBP instead of GIF? Here's how I do it. First, click "share" and choose the size you want. Then copy & paste that URL into a new tab (this part is key!). THEN, you can right click and "save photo" and you'll get the actual GIF.

171

Adam McCrea @adamlogic.com · 31/12/2024

We've been catching up our dependencies this month. Can you spot when we upgraded Postgres? 😲 We're on Timescale Cloud, upgraded from PG 13 to 17 (yeah, I know... 🤦‍♂️) CPU is way down, and our slowest query is down from 6s to 1s. Keep your dependencies up to date, people!!

1111

Adam McCrea @adamlogic.com · 30/12/2024

Question for the Ruby gem maintainers out there Do you commit your Gemfile.lock?... We do not, and it's so frustrating when CI fails because of a dependency update and NO changes in our code. Especially when it's been a month and LOTS of dependencies of changed. 😩

641

Adam McCrea @adamlogic.com · 20/12/2024

BONUS: Do you have multiple worker processes targeting different queues? Make sure you can tweak the concurrency of each process independently! But on Heroku you can't override an env var for a single process. The trick is to set it in the Procfile command itself.

020

Adam McCrea @adamlogic.com · 19/12/2024

Finally upgraded to Sidekiq 7. Can you spot it in this memory chart? Your eyes do not deceive you. It cut our memory usage IN HALF! I guess I can remove that "investigate high worker memory" issue from our backlog. 🥳

2142