This documentation explains how the Miren platform manages application scaling through automatic instance adjustment based on real-time traffic. It details the configuration of different scaling modes and the benefits of scaling services down to zero for resource efficiency.
Highlights
Supports two scaling modes: 'auto' for traffic-based adjustments and 'fixed' for maintaining a constant number of instances.
Features 'Scale to Zero' capability, which optimizes costs and resources by shutting down idle applications.
Configuration is managed via the .miren/app.toml file under the service concurrency settings.
Utilizes an observation-based model similar to Google Cloud Run rather than manual estimation.