forked from shudaniel/RuntimeAutoDeploy
-
Notifications
You must be signed in to change notification settings - Fork 0
Requirements
Aarti Jivrajani edited this page Jun 7, 2020
·
1 revision
- Provide an automated deployment mechanism that will spin up Kubernetes clusters and bootstrap an environment for user's applications to run in the cloud.
- Templatize the configuration files for deploying clusters in either JSON or YML files for users
User must specify:
- Repo URL
- This repo will be containerized and deployed.
- This repo will need to be containerized.
- Docker configuration file
- This file will be used to containerize the application and create the images.
- This file must also sit at the top level of the application.
- R.A.D. will containerize using docker and the user must specify a docker-compose.yml.
- Number of nodes
- Number of pods
- The user can specify which images will be deployed in which pods
- R.A.D. will provide an API endpoint to trigger setup and deployment of a cluster
- This API endpoint will take as input the configuration file
- Upon receipt of this file, the user can now manage the lifecycle of the clusters using the clusterctl CLI tool.
- We will use the Cluster API to handle deployment because it is a production-grade deployment package
- We will use "kind" to create local kubernetes clusters for development
- Cluster API provides the ability to create, scale, upgrade and destroy the cluster
- R.A.D. does not automate the management of an application
- The user is then responsible for managing everything using clusterctl