simple-pod-metrics-golang

Golang Simple Pod Metrics Example

This example shows how to make a Custom Pod Autoscaler (CPA) using Golang. The example extends the Alpine CPA base image (custompodautoscaler/alpine) and adds in a Golang binary to execute for gathering metrics/making evaluations. The code is verbosely commented and designed to be read and understood for building your own CPAs.

Overview

This example contains a docker image of the example Golang Custom Pod Autoscaler, alongside using the flask-metric sample application as a target to scale up and down.

Example Custom Pod Autoscaler

The CPA will try to ensure that there is always atleast 1 available across the resource and each app in the resource. The CPA will also ensure there are no more than 5 available across the resource. The CPA exposes two endpoints:

• GET /metrics
  • Displays all gathered metric values from every app instance in the resource.
  • Aliased with metrics in the example Dockerfile.
• GET /evaluation
  • Displays the evaluation decision made on how to scale, reporting the targetReplicas - how many replicas the resource should have.
  • Aliased with evaluation in the example Dockerfile.

Usage

Trying out this example requires a kubernetes cluster to try it out on, this guide will assume you are using Minikube.

Enable CPAs

Using this CPA requires CPAs to be enabled on your kubernetes cluster, follow this guide to set up CPAs on your cluster.

Switch to target the Minikube registry

Target the Minikube registry for building the image: eval $(minikube docker-env)

Deploy an app for the CPA to manage

You need to deploy an app for the CPA to manage:

• Build the example app image. docker build -t flask-metric ../flask-metric
• Deploy the app using a deployment. kubectl apply -f ../flask-metric/deployment.yaml Now you have an app running to manage scaling for.

Build CPA image

Once CPAs have been enabled on your cluster, you need to build this example, run these commands to build the example:

• Use the Makefile to build the docker image make docker
• Deploy the CPA using the image just built. kubectl apply -f cpa.yaml Now the CPA should be running on your cluster, managing the app we previously deployed.

Testing the CPA

• List pods. kubectl get pods -l app=flask-metric
• Exec into a pod. kubectl exec -it POD_NAME bash
• Get value. metric
• Increment value. increment
• Decrement value. decrement

Using this example as a starting point

This example uses a go.mod file to manage dependencies, you will need to set up a new go.mod file which uses your own package name. This example also uses a go.mod import that relies directly on the latest version of the custom pod autoscaler Go code, for your own autoscaler you would need to run:

go get github.com/jthomperoo/custom-pod-autoscaler/v2