Rails Autoscale

Simple, reliable queue-based autoscaling for your Rails app. Starting at $0/mo.

Rails Autoscale vs. Judoscale

If you’re deciding between Rails Autoscale and Judoscale, you can use either one since both add-ons are “Judoscale” behind the scenes. It’s a single product, but since we couldn’t rename the add-on, it has two faces in Heroku’s marketplace.

We recommend using Judoscale just so the branding you see in the app matches what you see in your Heroku add-on list. But really, it doesn’t matter. They are the same product, with the same pricing, supported by the same team. And they’re the best autoscaler out there.

Rails Autoscale vs. Heroku's autoscaler

Heroku has a native autoscaler available for free on performance dynos, so you might be wondering why you’d pay for another autoscaler. Here’s why:

  • Heroku only autoscales your web dynos. Rails Autoscale also autoscales your worker dynos running Sidekiq, Delayed Job, Resque, or Que.
  • Heroku autoscales based on total response time, often not improved by scaling. We use request queue time, which is a better metric for capacity and is almost always improved by scaling.
  • We autoscale your app over 3x faster than Heroku. This stops slowdowns before they become a problem, and you won’t get paged at 2am.
  • You can use Rails Autoscale with any dyno type that supports horizontal scaling.
  • Rails Autoscale gives you fine-grained control over your scaling behavior and a beautiful visualization UI.
A smarter way to autoscale

Rails Autoscale is the only autoscaler based on request queue time, not total response time. Some endpoints naturally take longer than others, and that shouldn’t trigger scaling. Rails Autoscale watches for real capacity issues to scale your app reliably and consistently.

Nate “If I was the king of the world, I would make it illegal to horizontally scale based on execution time. Scale based on queue depths, people!”
– Nate Berkopec, The Complete Guide to Rails Performance

Cut your Heroku bill in half

Most apps are overscaled so they can handle peak traffic. Rails Autoscale easily pays for itself by scaling down your app during quieter periods, Some apps earn back half of what they were previously paying for Heroku. Try it and be amazed at how many dynos you don’t need.

Mark “Rails Autoscale has allowed us to reduce our dyno usage for our primary application down to less than half of what we were using before, which is amazing (and allows us to not have to upgrade our Heroku plan). Thanks for making this, it really is a great little service.”
– Mark Urich

Impossibly simple

Rails Autoscale embraces the 80/20 rule and provides sensible defaults that work for 80% of apps. No sweat if you’re in the remaining 20%. Just a few tweaks is all it takes to get your app scaling smoothly.

Sam “Rails Autoscale saved me big time yesterday!! I’m new to this and I’m learning on the fly, but you made it very easy.”
– Sam Wood

Worker dynos included

Sidekiq, Resque, Delayed Job, and Que are all supported. Even if you split your worker queues across multiple processes, it’s no problem. Configure and autoscale each of them independently. Worker autoscaling is part of the package, no extra charge.

No API key required

Giving out your Heroku API key is no different than giving out your Heroku password. Rails Autoscale will never ask for it. Tight integration as a Heroku add-on means Rails Autoscale can only access the necessary endpoints of the Heroku API on behalf of your app, nothing more.

Extensive documentation

Rails Autoscale was built for Heroku, so the entire documentation site is designed to help you and your Heroku application. If you have a question, there’s a good chance we’ve answered it for you on the common questions page.

Not sure if Rails Autoscale is right for your app? Let’s chat.


Andrew “I’ve used most of them. Adam’s is the best. I highly recommend it.”
– Andrew Culver, Bullet Train

Region Availability

The available application locations for this add-on are shown below, and depend on whether the application is deployed to a Common Runtime region or Private Space. Learn More

  • Common Runtime
  • Private Spaces
Region Available
United States Available
Europe Available
Region Available Installable in Space
Virginia Available
Oregon Available
Frankfurt Available
Tokyo Available
Sydney Available
Dublin Available

Plans & Pricing

    • Max standard dynos across web & workers Unlimited
    • Max perf/private/shield dynos across web & workers Unlimited
    • Autoscale events per month 20
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
    • Max standard dynos across web & workers 5 dynos
    • Max perf/private/shield dynos across web & workers N/A
    • Autoscale events per month Unlimited
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
    • Max standard dynos across web & workers 10 dynos
    • Max perf/private/shield dynos across web & workers 3 dynos
    • Autoscale events per month Unlimited
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
    • Max standard dynos across web & workers 20 dynos
    • Max perf/private/shield dynos across web & workers 7 dynos
    • Autoscale events per month Unlimited
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
    • Max standard dynos across web & workers 50 dynos
    • Max perf/private/shield dynos across web & workers 15 dynos
    • Autoscale events per month Unlimited
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
    • Max standard dynos across web & workers 100 dynos
    • Max perf/private/shield dynos across web & workers 35 dynos
    • Autoscale events per month Unlimited
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
    • Max standard dynos across web & workers 250 dynos
    • Max perf/private/shield dynos across web & workers 80 dynos
    • Autoscale events per month Unlimited
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
    • Max standard dynos across web & workers Unlimited
    • Max perf/private/shield dynos across web & workers Unlimited
    • Autoscale events per month Unlimited
    • Autoscale based on request queue time
    • Autoscale worker dynos (Sidekiq, Resque, DJ, Que)
    • Precision autoscale controls
    • Scheduled autoscaling and pre-scaling
    • Live-updating metrics
    • Autoscale API
    • Developer support via email
Install Rails Autoscale
heroku addons:create rails-autoscale

To provision, copy the snippet into your CLI or use the install button above.

Rails Autoscale Documentation