Edwin Hernandez
Edwin Hernandez
> What are users supposed to do with that metric ? because it is essentially the startup latency of the model, it is useful in determining the pod autoscaling threshold...