Can someone tell me use cases that uses multiple combination of custom metrics to scale or combination of custom and system metrics?

when it comes to scaling in Kubernetes, we rely on metrics. By default, we use CPU and Memory metrics. Considering use cases where we need custom metrics (such as 'request count per second'), Kubernetes allows us to plug in custom metrics. For an application, there might be interrelated metrics that help in scaling. For example, combining "CPU" and "memory" metrics, we can use "CPU^2/memory" as the combined metric to scale. Can someone direct me towards real-world scenarios similar to this, but relating *custom metrics, where multiple *custom metrics are correlated in order to perform scaling? will this be a better approach?

Similar questions and discussions