Cloud Roadmap - Sept 2024 Team Meeting Notes

Goal: Get on the same page about our plan for OpenMRS supporting cloud hosting, multitenancy, and clustering. 

Context: Cloud support is one of our biggest deliverables for the TAP grant which ends Sept 29. Our main Cloud-related deliverable, "Create and share Cloud-Hosting Deployment Guidelines and Documentation including clustering guidelines", is rather behind schedule. We'd like to all get aligned on the plan for the next few weeks, and our cloud-support-related roadmap for the next year. 

This meeting was requested by a combo of Jan, Ian, Daniel, and Grace.

Notes

• Behind in our current schedule for deliverables around Cloud-hosting, but we can catch up! Part of the Scaleable workstream

  1. DIGI conversations / interviews for challenges in managing multiple implementations - wrapped up (link to notes/interviews/summary?)

  2. Cloud hosting Guidelines: Draft here: Cloud/Cluster Deployment

     1. to do

        1. how to deploy openmrs to kubernetes

        2. how to deploy instructions dbs and everythign that is needed in kubernetes

        3. terraform scripts, automation, whats needed in kubernetes to run openmrs

        4. not → how to run kubernetes, deploy container platform (can we link to that or mention how to rely on cloud host for it);

        5. ideally the intro would make it clear enough/basic/simple to know what to tell their technical team. Assume nothing → point to step by step instructions; Casey can be a good test person. Include Herbert in the read over along the way.

     2. need to update roadmap timeline

     3. Bahmni has similar setup and has about 80% that we could leverage - remaining estimated about 1 week’s worth of work

     4. Delayed to cybersecurity results - 1 left to request, + releases of modules; Raf unclear on timeline to complete this with Samuel L.

     5. Raf’s top priority will be the POC for the implementer’s meeting

  3. Sessions with community to explain to implementers devs on implementation of the guidelines

  4. O3 to FHIR store currently being generalized from CDI O3 implementation: CONNECTION TO HAPI FHIR SERVER - IMPLEMENTOR GUIDE

  5. Kubernetes vs Docker approach

     1. Start with one (Kubernetes), making sure it works and can be demo’ed

        1. Most cloud hosts standardize on Kubernetes

           1. is this too opinionated to only do kubernetes? maybe not, because of the standardization

           2. As long as we have OCI container images setup - best to have a default documented and supported way

        2. Then update for Docker Swarm post-Sept - or not, depending on if we want to take on the maintenance of this second documentation?

• Goal for Community Conference

  • Unconference session on “Cloud Hosting Support”

    • Herbert Yiga (UW DIGI) as host

    • Raff and Casey to join virtually (Grace bring Jabra speaker etc); Taib should be there

  • Share the link to the wiki page whenever cloud comes up in conversation: om.rs/cloud

  •  

Open