Feature Overview (aka. Goal Summary)

OpenShift workload partitioning allows to limit the cpu usage of all control plane components. MicroShift should be able to do the same. 

Goals (aka. expected user outcomes)

Requirements (aka. Acceptance Criteria):

1. Expose a configuration option to enable workload paritioning and configure the numbers of mcores allowed maximum for the control plane.
2. Re-Use OpenShift mechanismens to flag which components are considerd control plane components
3. Ensure all default components (network, storage etc) are flagged as control plane components.

Out of Scope

support low latency customer workload. see https://issues.redhat.com/browse/OCPSTRAT-361 for that.{}

Background

1. https://github.com/openshift/enhancements/blob/master/enhancements/workload-partitioning/management-workload-partitioning.md is the original design document describing the feature in single-node OCP.
2. https://github.com/openshift/enhancements/pull/1213 is a follow-up design extending support to multi-node clusters.
3. https://github.com/openshift/enhancements/blob/master/enhancements/shared-and-guaranteed-cpu-pools.md talks about splitting the CPU pool between "shared" and "guaranteed" segments.

Customer Considerations

none

Documentation Considerations

• Configuration option needs to be documented in the "configuring" book

Interoperability Considerations

None