Autoscaling for Heroku

Response times, queue times, job queues, load and schedule-based autoscaling for web and worker dynos

Features

Save Money

By shutting down idle resources when they're unnecessary, you can save money on your monthly bill.

Handle Load Fluctuation

Scale up and down automatically based on the load of your application using a variety of metrics, including job queues, queue times, and load.

Faster Job Processing

Spin up more workers than you usually would to process jobs at a faster rate without increasing your worker fees. Dyno fees are pro-rated to the second.

Configurable Scaling Frequency

Periodically perform checkups to determine whether your dyno formation needs to be adjusted with a base frequency of 60 seconds, all the way down to 15 seconds.

Scheduled Scaling

Optionally, or exclusively, scale based on a weekly schedule. Preemptively scale up before huge traffic spikes hit your application.

Heroku Logplex Integration

Integrating with Heroku Logplex allows your application to autoscale based on response times, queue times, requests per minute, and load.

Multiple Queues

Spread your workload out across multiple queues and dyno types, autoscale them individually. There are no limitations on the amount of Procfile entries and dyno managers that you can create.

Flexible App Integration

There are Ruby or Python libaries to get up and running with minimal effort. These libraries aren't mandatory and you can get setup easily without them as well. In some cases they aren't necessary at all.