Ona 3.6 Scoping Brainstorming

Owned by Craig Appl (Unlicensed)
Last updated: Feb 20, 2019

This document includes brainstorming notes from the Ona Team:

3 months --> December 5 -- March

  • Functional enhancements to Superset
    • Drilldown
    • Conditional formatting in tables
    • Annotations to bar charts
    • Drilldown
    • Pop-up with new indicator
    • Legend
    • Admin ability to upload shapefiles more easily
    • Build a new user model specifically for OLMIS to make sure that within Superset the user model is queried instead of having the WHERE clause in the indicator
      • Not sure if WHERE clause is secure
      • Test whether someone with VIEW permissions can hack it and remove the WHERE class
      • Do we need to restrict someone's edit permissions to their current role?
      • Superset is built on FAB --> FAB doc'n talks about extending user model to collect other fields
    • PDF
      • Headers/footers/logos
      • Chart
      • --> modify CSS to dynamically display what dashboard should look like
    • XLSX
      • Add nice logos and headers to CSV
      • Specifically say what the column is supposed to be
      • Add footers (metadata of export)
    • Make a dashboard responsive based on browser's size
    • Create a print layout --> media query --> when you press Print button you can define size of thing you want to print and specifically adjust layout to be formatted appropriately for the page
    • Align look and feel of Superset dashboards with OLMIS style guide
    • Visualization
    • Map-specific
    • Dashboard-level access controls
    • Row-level filtering
    • Exports
    • Modify CSS
  • Stock management reporting in Postgres for 3.6?
  • SSO
    • Improvement around when login occurs
  • Kafka
    • Develop a use case that plays to Kafka's strengths
    • Test it
    • Document it
  • Data pumps
    • Debezium - The reporting team uses the API and extracts it. This tool allows us to let the service owner define what gets sent to the reporting stack.
      • Instead of fetching the data, we would be able to receive it as it happens.
      • Can we fit this in with DHIS2?
  • DHIS2
    • Make reports available or post them to DHIS2
  • DevOps
    • Clearly identify DevOps requirements with Peter
    • Reporting stack we've built have been built for developer use, not production use
    • Persist data
    • Figure out how to upgrade each individual service in a way that retains the data and possibly migrates it
      • Biggest risk is upgrading NiFi
      • Kafka
    • NiFi auto load sensitive values
    • Superset auto load custom user roles
    • Superset auto load dashboards + charts
  • Knowledge transfer to SolDevelo and JSI teams
    • In-person meeting
    • Define process for adding new microservices to reporting stack
  • NiFi registry and RBAC
  • Add a "retrieve all" token that returns all requisitions regardless of the user's permissions

Additional Meeting Notes from meeting on 20 Nov 2018

Attendees:

  • Clay
  • Craig
  • Josh
  • Mary Jo
  • Wes

Discussion:

  • Attendee Priorities:
    • Josh: DHIS2 and devops only if it makes it easier for new teams to learn. We need to make it useful to deploy for developers in production.
    • Clay: Devops and improvement of SSO
    • Craig: Devops and Superset row level filtering/permissions
    • Wes: DHIS2, exports from Superset (power users need to have access to the data) and visualizations. We need to build capacity of the team.
    • Mary Jo:
      • Musts:
        • Knowledge Transfer - will be meeting in person
        • DHIS2
      • Priority:
        • Superset enhancements (TBD)
  • Next:
    • Scope DHIS2
      • Decide on the scope and there are really only two pieces of work:
        • Make KPIs
        • Move them
    • Create Level of Effort of things we can't get
    • Create an epic for each with the Component as "Reporting"
    • To prepare for Zambia:
      • Who is going to be working on what?
      • They want a holistic list of what is in OpenLMIS v3.6
    • Review the Epics that are in Reporting and mark them dead

Further Discussion 20 Feb 2019

Attendees: Wesley Brown, Clay Crosby (Unlicensed), Craig Appl (Unlicensed)

Discussion:

  • There's a risk in the FHIR approach that Craig is bringing up with Josh
  • It looks like BAO will do the DHIS2 transportation, so we will descope Ona's role on that.
  • Ona will be responsible for providing an interface for the reporting solution using a standard
  • Things that need to be done should be done by sprint 121 - 27th March
  • Production Readiness is nearly done
  • Persisting information in Superset is done
  • Decisions:
    • CSS Modifications
    • Print PDFs
    • Nifi Security is too big with a small value add
    • Continue with building standard interface that can be consumed by DHIS2

OpenLMIS: the global initiative for powerful LMIS software