HireFire provides the ability to auto-scale your web- and worker dynos. It supports a variety of metric types. These metric types are used as a unit of measurement to determine, in combination with your HireFire configuration, when and how to scale.

For web-based dynos we support the following metrics types:

  • Response Time (percentile, average)
  • Connect Time (percentile, average)
  • Queue Time (percentile, average)
  • Dyno Load (average)
  • Requests Per Minute
  • Apdex Score

For worker-based dynos we support the following metric types:

  • Job Queue
  • Dyno Load (average)

Guides

We've prepared a few handy guides that'll walk you through integrating HireFire with your web-type dynos. We personally recommend the Logplex solutions over the (legacy) HireFire and New Relic solutions. This collection also includes guides on how to integrate HireFire for Job Queue-based scaling for worker-type dynos various environments.

Integration

While there are libraries for Ruby and Python, you don't necessarily require any of these. They are simply provided for convenience and are only useful when you're looking setup Job Queue-based auto-scaling. All of the other metric types mentioned above won't involve any libraries at all and are easy and quick to
setup. 

Did this answer your question?