OCPDAS - Supporting Code

Supporting code for the paper "OCPDAS: Online Change-Point-Detection-Based Autoscaling for Stable Pod Scaling in Kubernetes".

Environment

Docker

We recommend installing Docker Engine following the official documentation.

Minikube

We recommend installing Minikube, a local Kubernetes cluster, following the official documentation. Then start Minikube by running:

minikube start --driver=docker --cni=bridge
minikube addons enable metrics-server

Kubectl

We recommend install kubectl following the official document. Then create an independent namespace named ocpdas for testing the repository by running:

kubectl create namespace ocpdas

all the benchmark tests are conducted in the ocpdas namespace and you can safely remove all content after reproduction the namespace following

kubectl delete namespace ocpdas

Conda Environment

We recommend using conda to manage the running environment to reproduce the experiments.

conda env create -f environment.yml

Reproduction

Follow the instructions in the README.md file in every subfolder to run the code.