Idle web container continues accruing container-hour charges at zero traffic

Can Scalingo web containers scale to zero?

Scalingos autoscaling model for web processes typically enforces a minimum of one running instance. Setting the minimum to zero is not supported in the current platform configuration, meaning the system will always keep at least one container active to handle incoming requests.

If you manually scale the web process to zero via the CLI or dashboard, the application becomes unreachable. No automatic restart occurs when traffic arrives later; a new container must be started explicitly.

Regarding costs, add-on services such as databases, caching, and monitoring are billed independently of the web container count. Their charges continue to accrue even when all web containers are stopped.

1. Run scalingo scale web=0 and observe whether the process stops and stays stopped.
2. Send a request to your apps URL afterward; expect a 503 or connection error if no container is running.
3. Check your billing dashboard or CLI output for add-on charges while the web process is at zero.
4. Review autoscaling settings with scalingo autoscale --list to confirm the minimum instance setting.

Note: Autoscaling policy and billing behavior may differ between Scalingo plans (e.g., free vs paid tiers) or after recent platform updates. The information here reflects the model as of 19 October 2026.

If you are unsure whether your plan supports a zero-minimum configuration, please let me know your plan name so I can tailor the next steps.