j4ckstraw

[question] ClusterColocationProfile是否考虑支持priority和只匹配一次？

1

**What happened:** 我们在内部灰度koordinator方案，针对不同的作业预制了多个ClusterColocationProfile，通过命名空间级别来灰度。 但是命名空间对于业务来说可能过大了，需要更精细的灰度方案，所以定制了多个ClusterColocationProfile，但是发现不同的ClusterColocationProfile会相互覆盖，从[代码](https://github.com/koordinator-sh/koordinator/blob/9a76018e38730b8595a09a563b91c43abdb31423/pkg/webhook/pod/mutating/cluster_colocation_profile.go#L89)看也是支持匹配多个profile的，想了解一下如此设计的背景和目的。 **What you expected to happen:** 希望增加优先级支持和只匹配一次的能力。 **Environment:** - Koordinator version: - v1.3.0 - Kubernetes version (use kubectl version): v1.21.3 - docker/containerd version: containerd 1.5.0 - OS...

[proposal] call eviction API before kill container

1

**What is your proposal**: call eviction API before kill container **Why is this needed**: memoryEvict will kill container then call eviction api, this may disobey pod disruption budgets (PDB) https://github.com/koordinator-sh/koordinator/blob/661dd71df888a7cb8e2b0891408b34225dc2ee85/pkg/koordlet/qosmanager/plugins/memoryevict/memory_evict.go#L163...

1. communicate to kube-apiserver with protobuf 2. listOption add resourceVersion=0. without resourceversion, list will force kube-apiserver retrieve data from etcd. In a 100+ nodes, 7500+ pods kubernetes cluster, this patch...

*Issue number of the reported bug or feature request: #* **Describe your changes** A clear and concise description of the changes you have made. do not ping not ready goldpinger...

[proposal] make use of batch/mid resource of node capacity

3

**What is your proposal**: Change capacity of batch/mid resource to node capacity * reclaimThreshold to make use of capacity field. https://github.com/koordinator-sh/koordinator/blob/main/pkg/slo-controller/noderesource/plugins/batchresource/plugin.go#L147 **Why is this needed**: **Is there a suggested solution,...

koordlet: add mid runtimehooks

5

### Ⅰ. Describe what this PR does ### Ⅱ. Does this pull request fix one issue? fix https://github.com/koordinator-sh/koordinator/issues/1442 ### Ⅲ. Describe how to verify it ### Ⅳ. Special notes for...

it will be better to get pods from kube-apiserver cache with resourceVersion="0" see https://github.com/kubernetes/kubernetes/issues/102672#issuecomment-1026197852

feat: support change pod CPUQOS by annotations

2

### Ⅰ. Describe what this PR does ### Ⅱ. Does this pull request fix one issue? ### Ⅲ. Describe how to verify it ### Ⅳ. Special notes for reviews ###...

**What happened**: Take koordlet as example, because it import `net/http/pprof`, and main goroutine always listen on addr which default is :9316, so we can always fetch pprof from :9316/debug/pprof/xxx **What...