Judoscale is a dedicated autoscaling service that helps engineering teams manage application capacity across Heroku, Render, Railway, Fly.io, Scalingo, and Amazon ECS. Rebranded from Rails Autoscale in 2022, it supports applications built with Ruby, Python, Node.js, and Java, including web frameworks and job queues.

The service uses request and job queue time to determine when capacity should increase or decrease. Customers control minimum and maximum instance counts, the number of instances changed per scaling event, scaling frequency, and schedules for predictable traffic. Judoscale’s algorithm checks every 10 seconds and can scale both web processes and asynchronous workers to help maintain throughput during demand spikes.

By adjusting resources to match actual capacity needs, Judoscale aims to reduce overprovisioning and hosting costs while helping prevent slowdowns and backed-up queues. The company says its language-specific packages collect queue-related metrics and basic process metadata, not request or job contents. Judoscale offers a free trial without requiring a credit card, and provides documentation, platform-specific pricing, and consultation options.