Current Thinking in Reporting Stack First Run

Last Updated 18 July 2018

This document defines our current thinking and approach for standing up the reporting stack first run experience.

Current Status:

• We have created docker containers for:
  • Nifi
  • Postgres
  • Superset
  • Kafka
  • Zookeeper
  • Nifi Registry
• We have created templates for:
  
  • Creating a database connection pool - The database currently needs to be created outside of Nifi
  • Creating a table in Postgres from Nifi
  • Connecting to Superset to create a datasource in the API
  • A number of processor groups to extract data from the OpenLMIS stock management endpoints

Next Priorities:

1. Automate the process for creating the database schema - We still haven't been able to define which schema to use (openlmis v2 or v3)
2. Automate the process for connecting to the Nifi Registry
3. Stand up a Nifi Registry as a service within OpenLMIS infrastructure
4. Build metrics in Superset
5. Integrate the process with Jenkins

Timeline:

We hope to have each of these steps complete during the next sprint and tested. Beyond that, we need to work on the following

• Release the reporting stack
• User Credentials for Nifi flows
• Figure out the new Oauth controller service against the OpenLMIS Oauth process