Horizontal worker dyno scaling in Scalingo v2.0 lacks auto-scaling thresholds

Scaling Capabilities in Scalingo v2.0

Scalingo v2.0 does not support configurable auto-scaling thresholds for worker dynos. Horizontal scaling for worker dynos remains a manual-only operation.

Current Scaling Behavior

Based on the v2.0 architecture, the platform utilizes a static scaling model. There are no native mechanisms to define triggers based on CPU utilization, memory consumption, or request volume. Consequently, the scaling API is designed to report the current state of deployed dynos rather than exposing or managing policy-driven rules.

Confirmed Facts

• Manual Control: Scaling must be initiated via the CLI or API by specifying the exact number of desired dynos.
• Scope: Horizontal scaling is restricted to worker dynos; web dynos operate under a separate scaling logic.
• Configuration: Scaling overrides are handled through static environment variables or application settings, not dynamic thresholds.
• API Output: The scaling API returns integer counts of active dynos and does not include policy objects or pending auto-scale events.

Implementation Steps for Scaling

To adjust the capacity of your worker dynos, use the following manual approach:

1. Invoke the scaling command via the CLI to set the specific dyno count.
2. Verify the change through the scaling API to confirm the current count.
3. Monitor application logs to ensure the new dynos have successfully initialized.

Technical Considerations

When performing manual scaling, be aware of two primary constraints:

• Caching Delays: CLI output may not reflect the change immediately due to known caching delays.
• Regional Caps: Scaling requests may fail if the target count exceeds region-specific resource limits, even if the API accepts the initial command.

Diagnostic Requirement: To provide a more specific recommendation for automating your workload, please specify if you are using an external monitoring tool to trigger these API calls.