add memory qos strategy

[saintube]

Signed-off-by: saintube saintube@foxmail.com

Ⅰ. Describe what this PR does

add Memory QoS strategy to enhance resource assurance for k8s pods.

1. MinLimtPercent: allow pods to lock some memory cache from global reclaim.
2. ThrottlingPercent: allow pods to throttle memory allocation before reaching the memory limit.
3. WmarkRatio: allow pods to get memory reclaimed asynchronously before direct reclaim.
4. WmarkMinAdj: allow a node to prioritize global reclaim and make less disruption for high-priority pods.

Some of the above abilities require a kernel of cloud kernel/ Open Anolis OS.

related work: kubernetes/enhancements#2571

Ⅱ. Does this pull request fix one issue?

fixes #55

Ⅲ. Describe how to verify it

1. Enable feature-gate CgroupReconcile in Koordlet.
2. Turn on MemoryQOS in configmap slo-controller-config.
3. Try colocation.

e.g.

apiVersion: v1
data:
  resource-qos-config: |-
    {
      "clusterStrategy": {
        "lsClass": {
           "memoryQOS": {
             "enable": true
           }
         }
      }
    }
kind: ConfigMap
metadata:
  name: slo-controller-config
  namespace: koordinator-system

Ⅳ. Special notes for reviews

None

[codecov-commenter]

Codecov Report

Merging #101 (cbae9d2) into main (ccbad0c) will increase coverage by 3.13%.
The diff coverage is 70.71%.

❗ Current head cbae9d2 differs from pull request most recent head 24fb108. Consider uploading reports for the commit 24fb108 to get more accurate results

@@            Coverage Diff             @@
##             main     #101      +/-   ##
==========================================
+ Coverage   44.19%   47.32%   +3.13%     
==========================================
  Files          73       75       +2     
  Lines        5965     6658     +693     
==========================================
+ Hits         2636     3151     +515     
- Misses       3036     3192     +156     
- Partials      293      315      +22     

Flag	Coverage Δ
unittests	47.32% <70.71%> (+3.13%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
pkg/koordlet/resmanager/config.go	58.33% <ø> (+4.48%)	⬆️
pkg/koordlet/resmanager/resmanager.go	2.75% <0.00%> (+0.02%)	⬆️
pkg/util/config.go	0.00% <0.00%> (ø)
pkg/util/container.go	25.00% <0.00%> (-1.82%)	⬇️
pkg/util/pod.go	10.44% <0.00%> (-0.58%)	⬇️
pkg/util/system/cgroup.go	70.00% <0.00%> (ø)
pkg/util/system/cgroup_resource.go	0.00% <ø> (ø)
pkg/slo-controller/nodeslo/nodeslo_controller.go	60.15% <42.85%> (-0.97%)	⬇️
pkg/koordlet/resmanager/types.go	57.57% <64.70%> (+16.03%)	⬆️
...kg/koordlet/resmanager/resource_update_executor.go	66.89% <80.82%> (+23.84%)	⬆️
... and 7 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ccbad0c...24fb108. Read the comment docs.

[jasonliu747]

/lgtm

[zwzhang0107]

/lgtm

[hormes]

/approve

[koordinator-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: hormes

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

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [hormes]

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