Background
In preparation for the EBS V3 release, the GOBii-EBS team has started working on refactoring code, creating the pipeline to build new EBS-GOBii containers, and a test swarm cluster – all as preliminary work for the phase 1 of GOBii-EBS integration. Although the foundation has already been laid down through the 2019 EBS DevOps Hackathon, a lot of improvements have been identified and must be implemented. This product iteration’s goal is to have an initial build of the EBS-GOBii containers that satisfy parts of this architecture:
Goals
For this sprint, the focus is on putting the infrastructure, particularly automation, in place for the team to be able to develop software efficiently and conform to the Agile methodology. This will also include some preliminary work on the middleware to prepare it for the automated data-loading which will be part of EBS v3. By the end of this sprint, the following should be available:
Separate code branches for all GOBii git repositories for the following products: GOBii-standalone and EBS-GOBii (refer to Git Branching Strategy for the technical details of this)
GOBii-standalone CI/CD pipeline pointed to the branch GOBii-standalone (called “standalone_develop”)
EBS-GOBii CI/CD pipeline that will automatically build the EBS-GOBii containers and create the necessary artifacts and release packages - based off of the EBS-GOBii branch (called “ebs_develop”)
Test Swarm Cluster in CBSU servers for the GOBii-EBS pipeline to be able to deploy containers to
Dockerhub repositories that will hold the new docker images
Docker swarm stack file that will be able to deploy the new docker containers to a swarm cluster
A process (and infrastructure) in place to be able to do product increment demonstration every end of sprint (during the sprint review meeting)
Instruction objects in place of instruction files - in preparation for the automated data loading functionality
EBS-GOBii:
Test swarm cluster (4-node) and staging swarm cluster (11-node)
Docker swarm support
EBS-GOBii CD
Testing the MonetDB-free system
Initial integration with the Service Gateway
Finish refactoring of the middleware (aspect files, Digesters)
Decoupling of Timescope with the web container
Postgres 12
JIRA
Fix Version: GE_20.6
Questions
Below is a list of questions to be addressed as a result of this requirements document:
Question | Outcome |
|---|---|
Assumptions
Mockups
What we're not Doing
Integration of database deployment processes to the DevOps pipeline.
Standardizing the continuous integration (CI) practices across EBS software development teams. Although this is something we would like to do in future sprints, for now the only requirement is software artifacts are to be released as Docker images.