The reporting stack has been developed to make it easy for developers to spin up the stack, develop reports, export them and setup demonstration systems. The reporting stack should be structured differently when deployed on hardware in production environments because implementations need to account for autoscaling, zookeeper management and running without docker.
This epic focuses on assessing the differences between the current reporting stack and the recommended production deployment tooling, defining a list of items to improve and identifying if we want to work on them for the OpenLMIS project.
Evaluate and document best practices for running each component in a production system
Develop a mechanism to persist data in all systems
Identify a process for upgrading each service to stay with the current tech
Determine a mechanism to store sensitive values and make them available in production environments