Querying the auto-scaling formula

[clxseizetheday]

hello guys, I am reading the document about the autoscaling of OpenFaas. In the chapter 'Autoscaling', https://docs.openfaas.com/architecture/autoscaling/#how-auto-scaling-works it gives a formula about the strategy of calculating the desire replics amount.

it says desired = current replicas * (current load / (target load per replica * current replicas)) however, the first current replicas in the formula and the current replicas in the denominator will always cancel out. So how should I understand this formula?

Thanks

[clxseizetheday]

In fact, if I cancel out the two current replicas in the formula, it will turn into desired = current load / target load per replica. This new formula is easy to understand. In my opinion, current load will not change before and after the sacling, so just divide target load per replica by current load and then you get the desired replics, right? How does current replicas influence the origin formula?

[LucasRoesler]

@alexellis I think it might be best for you to comment here, given your recent work on scaling

[alexellis]

The formula should match that of Kubernetes HPAv2:

https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale/#algorithm-details

[alexellis]

/set title: Querying the auto-scaling formula