Scalingo web container autoscaling metric target and cooldown configuration

Direct answer

Current model knowledge does not contain reliably confirmed current defaults for Scalingo web container autoscaling aggregation window, target interpretation per-container vs total, and scale-up / scale-down cooldown values. Platform documentation is version-sensitive and the exact semantics require current verification in the Scalingo dashboard and official docs for your service plan.

Confirmed from available knowledge: min and max container bounds are configurable for a Scalingo service and act as hard limits for autoscaling, independent of metric target and aggregation settings.

What can be verified now

• Min / max containers: configurable and enforced.
• Metric choice: RPM or resource usage can be selected as the autoscaling driver.
• Target value and aggregation window are part of the autoscaling configuration UI.

Likely explanation vs confirmed facts

Unconfirmed, requires verification:

• Aggregation window applied to the chosen metric when evaluating the target. Model knowledge does not establish a current default window for Scalingo.
• Target interpretation: per-container average vs total across containers. PaaS platforms commonly describe resource usage targets as per-container average and request rate targets as total or per-container depending on implementation, but confirmed Scalingo behavior for RPM or resource usage metrics cannot be verified from model knowledge alone.
• Cooldown configurability: whether Scalingo allows separate scale-up and scale-down cooldowns versus a single shared cooldown, and what the current default values are, is not reliably established in model knowledge and is subject to platform changes.

Confirmed: Min and max bounds are configurable hard limits.

Steps to get a definitive answer for this case

1. Check Scalingo official documentation Autoscaling section for current metric aggregation window, target semantics, and cooldown parameters.
2. Inspect Scalingo dashboard Autoscaling configuration UI for the specific service to see configured window, target type, and cooldown settings.
3. Review scaling events log in Scalingo dashboard to observe actual evaluation intervals and cooldown enforcement in practice.

One missing diagnostic detail that changes the recommendation: which metric type you are configuring, RPM vs CPU / memory usage, and the service plan. Target interpretation and sensible window / cooldown choices differ between request-rate and resource metrics.