WIP: Added support for performance profile status

[rbaturov]

This PR includes the implementation of this PR on the hypershift side:
openshift/enhancements#1619

Up until now, there was no defined resolution over how PerformanceProfile
status conditions with the status of the different components created
from the PerformanceProfile (MachineConfig, KubeletConfig, Tuned) are handled and exposed.
To resolve this, a custom ConfigMap is added and will be used as a middleware object.
It will hold the updated status populated by the performance profile controller
and will be watched by the NodePool controller, which will calculate an overview
of the current status provided and reflect it under NodePool.status.condition.

The resolution described above consist of the following changes:

• Added a new condition "PerformanceProfileAppliedSuccessfully" to the NodePool conditions API.
• Added a watch to the specific performance profile status ConfigMap object that will trigger the reconcile loop.
• Added the HandlePerformanceProfileStatus function to calculate and set the PerformanceProfileAppliedSuccessfully based on the possible performance profile statuses (available, progressing and degraded).

[openshift-ci]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: rbaturov
Once this PR has been reviewed and has the lgtm label, please assign bryan-cox for approval. For more information see the Kubernetes Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[openshift-ci]

@rbaturov: The following tests failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
ci/prow/verify	9bc1534	link	true	/test verify
ci/prow/e2e-kubevirt-aws-ovn	9bc1534	link	true	/test e2e-kubevirt-aws-ovn

Full PR test history. Your PR dashboard.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.

[stevekuznetsov]

Where does the node tuning operator usually put this information? Can the NodePool controller read it directly from there, instead of using a middleware object?

[rbaturov]

Where does the node tuning operator usually put this information? Can the NodePool controller read it directly from there, instead of using a middleware object?

In NTO on OCP, we have performanceprofile object and the performance profile controller is responsible for writing this information to the object status.
However, on hypershift performanceprofiles are applied as configmaps. Thus, adjustments for this are required in the NTO.
Since NTO with these adustements can't change the ConfigMap and not the PerformanceProfile embedded in it since it controlled by the Hypershift operator which will override the changes immediately.
The Nodepool is the object which suppose to report about the status of the PerformanceProfile since it represents the current configuration of the nodes.
Since NTO cannot modify the Nodepool directly, a middle-ware object (ConfigMap) is used to write to it the status and the nodepool controller will read from it.

[stevekuznetsov]

Sorry if these are obvious questions, but I am new - using proxy objects to funnel state in a distributed system can have a lot of downsides w.r.t. consistency, especially around outages, as etcd has no transactions and no way to keep objects in sync.

However, on hypershift performanceprofiles are applied as configmaps. Thus, adjustments for this are required in the NTO.

Are performanceprofiles cluster-scoped in OCP core? Have we done a spike to see if we could namespace them and re-use that object validation for HCP?

Since NTO cannot modify the Nodepool directly,

Why?

[rbaturov]

Sorry if these are obvious questions, but I am new - using proxy objects to funnel state in a distributed system can have a lot of downsides w.r.t. consistency, especially around outages, as etcd has no transactions and no way to keep objects in sync.

However, on hypershift performanceprofiles are applied as configmaps. Thus, adjustments for this are required in the NTO.

Are performanceprofiles cluster-scoped in OCP core? Have we done a spike to see if we could namespace them and re-use that object validation for HCP?

Yes performance profile are cluster scoped.

Since NTO cannot modify the Nodepool directly,

Why?

He doesn't have RBAC permissions for that