cpureplicaprint

CPU Replica Print

This example shows how the library can be used to both gather metrics based on metric specs, and then calculate the replica count that the Horizontal Pod Autoscaler (HPA) would target based on those metrics.

In this example a deployment called php-apache is created with 4 replicas that responds to simple HTTP requests with an OK!. The example will query the CPU metrics for the pods in this deployment, along with the number of replicas the HPA would target based on those metrics and print them to stdout.

Note this example uses out of cluster configuration of the Kubernetes client, if you want to run this inside the cluster you should use in cluster configuration.

Usage

To follow the steps below and to see this example in action you need the following installed:

• Go v1.22+
• K3D v5.6+

After you have installed the above you can provision a development Kubernetes cluster by running:

k3d cluster create

Steps

Run go get to make sure you have all of the dependencies for running the application installed.

1. First create the deployment to monitor by applying the deployment YAML:

kubectl apply -f deploy.yaml

1. Run the example using:

go run main.go

1. If you see some errors like this:

2022/05/08 22:26:09 invalid metrics (1 invalid out of 1), first error is: failed to get resource metric: unable to get metrics for resource cpu: no metrics returned from resource metrics API

Leave it for a minute or two to let the deployment being targeted (php-apache) to generate some CPU metrics with the metrics server.

Eventually it should provide output like this:

2022/05/08 22:27:39 CPU statistics:
2022/05/08 22:27:39 Pod: php-apache-d4cf67d68-s9w2g, CPU usage: 1m (0.50% of requested)
2022/05/08 22:27:39 Pod: php-apache-d4cf67d68-v9fc2, CPU usage: 1m (0.50% of requested)
2022/05/08 22:27:39 Pod: php-apache-d4cf67d68-h4z4k, CPU usage: 1m (0.50% of requested)
2022/05/08 22:27:39 Pod: php-apache-d4cf67d68-jrskj, CPU usage: 1m (0.50% of requested)
2022/05/08 22:27:39 Based on the CPU of the pods the Horizontal Pod Autoscaler would scale from 4 to 0 replicas
2022/05/08 22:27:39 ----------

1. Try increasing the CPU load:

kubectl run -it --rm load-generator --image=busybox -- /bin/sh

Once it has loaded, run this command to increase CPU load:

while true; do wget -q -O- http://php-apache.default.svc.cluster.local; done