Introduce cluster id for MicroShift

Feature Overview (aka. Goal Summary)

MicroShift cluster instance should have a unique cluster-id, like OpenShift clusters. Just relying on RHEL UUID or HostName/IP adresses is not "unique" enough 

Goals (aka. expected user outcomes)

• be able to uniquely identify and distinguish MicroShift clusters
• We need to be able to identify/distinguish different clusters, e.g. for support cases, fleet management, telemetry etc. 

Requirements (aka. Acceptance Criteria):

1. Upon initial installation/startup, a globally unique ID shall be created.
2. This ID shall not be modified/changed throughout the lifetime of the cluster (e.g. during upgrades/rollback)
3. Special care must be taken for ostree based installation - if the same image is instantiated multiple times on multiple devices, the resulting cluster must have different IDs (meaning the cluster ID must not be created during image generation, but at initial installation/startup).
4. Document ways for the end-users to get the cluster ID information 
5. The cluster ID must be part of every sosreport.
6. The cluster ID shall not be "easily" changeable by users. 
7. Explore how OCP exposes the cluster ID and decide if we can do it in the same way in MicroShift (if feasible), because tht would make mmgt tools like ACM, ACS etc which show/use a clusterid work unmodified with MicroShift cluster-id 

Out of Scope

n/a

Background

n/a

Customer Considerations

n/a, as this is intended for internal use only

Documentation Considerations

• In the "Support" book, there should be a chapter "Obtaining your cluster ID"

Interoperability Considerations

None