Feature Overview (aka. Goal Summary)

Reduce the resource footprint of LVMS in regards to CPU, Memory, Image count and size by collapsing current various containers and deployments into a small number of highly integrated ones.

Goals (aka. expected user outcomes)

• Reduce the resource footprint while keeping the same functionallity
• Provide seemless migration for existing customers{}

Requirements (aka. Acceptance Criteria):

• Ressource Reduction:
  • Reduce memory consumption
  • Reduce CPU consumption (stressed, idle, requested)
  • Reduce container count (to reduce APi/scheduler/crio load)
  • Reduce container image sizes (to speed up deployments)
• new version must be functionally equivalent no current feature/function is dropped,
• day1  operations (installation, configuration) must be the same
• day2 operations (updates, config changes, monitoring) must be the same
• seamless migration for existing customers
• special care must be taken with MicroShift, as it uses LVMS in a special way.

Questions to Answer (Optional):

1. Do we need some sort of DP/TP release, make it an opt in feature for customers to try? 

Out of Scope

tbd

Background

The idea was creating during a ShiftWeek project. Potential saving / reductions are documented here: https://docs.google.com/presentation/d/1j646hJDVNefFfy1Z7glYx5sNBnSZymDjCbUQVOZJ8CE/edit#slide=id.gdbe984d017_0_0 

Customer Considerations

seamless migration for existing customers

Documentation Considerations

Resource Requirements should be added/updated to the documentation in the requirements section.

Interoperability Considerations

Interoperability with MicroShift is a challenge, as it allows way more detailed configuration of topolvm with direct access to lvmd.conf