Version:

Use external scalers

This tutorial shows how to use an external resource scaler, for example, HorizontalPodAutoscaler (HPA) or Keda's ScaledObject, with the Serverless Function.

Keep in mind that the Serverless Functions implement the scale subresource, which means that you can use any Kubernetes-based scaler.

Prerequisites

Before you start, make sure you have these tools installed:

Kyma installed on a cluster

Steps

Follow these steps:

HPA
Keda CPU
Keda Prometheus

Create your Function with the replicas value set to 1, to prevent the internal Serverless HPA creation:

Click to copy

cat <<EOF | kubectl apply -f -
apiVersion: serverless.kyma-project.io/v1alpha2
kind: Function
metadata:
  name: scaled-function
spec:
  runtime: nodejs16
  replicas: 1
  source:
    inline:
      dependencies: ""
      source: |
        module.exports = {
          main: function(event, context) {
            return 'Hello World!'
          }
        }
EOF

Create your HPA using kubectl:

Click to copy

kubectl autoscale function scaled-function --cpu-percent=50 --min=5 --max=10

After a few seconds your HPA should be up to date and contain information about the actual replicas:

Click to copy

kubectl get hpa scaled-function

You should get a result similar to this example:

Click to copy

NAME              REFERENCE                  TARGETS   MINPODS   MAXPODS   REPLICAS   AGE
scaled-function   Function/scaled-function   1%/50%    5         10        5          61s