We scale up your dynos immediately when queue time exceeds this range. When queue time settles below this range, we’ll scale your dynos down.

You can scale up by multiple dynos at a time to aggressively recover from a slowdown.

Minimum duration to wait before subsequent upscale events. Shorter frequency will scale up faster.

Increase the sensitivity to avoid upscales during a single momentary spike, waiting for a more sustained slowdown before scaling up.

You can scale down by multiple dynos at a time to optimize for cost.

Shorten the time between downscales to optimize cost savings. A longer delay avoids premature downscaling.