Skip to content

Latest commit

 

History

History
28 lines (22 loc) · 2.73 KB

File metadata and controls

28 lines (22 loc) · 2.73 KB

Kubernetes-based Event Driven Autoscaling with KEDA - A Practical Guide

This reposiory contains source code for the article Kubernetes-based Event Driven Autoscaling with KEDA- A Practical Guide

In the article, the basics of Kubernetes Event Driven Autoscaling (KEDA) are first explained. After this is described how to install a lightweight local K8S cluster called Minikube, in which KEDA is also installed. To then demonstrate how KEDA works in practice, a use case is considered for dynamically scaling based on a certain amount of .csv-files that need processing. After following the guide, a person should be able to get KEDA up and running and have a good understanding of the core concepts of it.

Getting Started

To get started with KEDA in practice, simply follow the steps that are outlined in the article.

Repository Structure

  • The folder manifests contains a number of manifests for deploying K8S resources into the Minikube cluster:
    • deployments/redis-deployment.yaml: Contains definition for deploying a Redis server used as a dynamic scaling trigger.
    • helpers/pvc-inspection.yaml: Contains definition for an inspection pod
    • jobs/csv-processor-scaled-jobs.yaml: Contains the definition of ScaledJob used for aggregating CSV files. Central file for KEDA
    • jobs/data-generator.yaml: Contains definition for a job that generates raw CSV files that should be aggregated by the ScaledJob processes
    • services/redis-service.yaml: Contains definition for a Redis service that exposes the Redis deployment within the cluster
    • volumes/data-pvc.yaml: Contains definition for a Persitent Volume Claim used for storing and reading CSV data
  • The folder src contains two .py-files that are used for creating raw CSV data and for aggregating the CSV data:
    • src/data_generator.py: Contains the code needed for generating the CSV files. Is used in the data-generator definition via the data_generator.Dockerfile image
    • src/process_csv.py: Contains the code needed for aggregating CSV files. Used in the ScaledJob definition via the csv_processor.Dockerfile image
  • The root folder contains the definitions of the data_generator.Dockerfile and csv_processor.Dockerfile, which are used in the jobs

Contributing

We welcome contributions from the community! If you find a bug or would like to suggest a new feature, please create a Github issue and we'll take a look.

About the Author

The article was written by Tim Molleman, a Data Engineer at Digital Power

If you're interested in learning more about how we use other cloud and its services to build scalable, resilient solutions for our customers, please check out our website.