kubectl-view-allocations

kubectl 插件列出节点和运行中的 Pod 清单中定义的资源（CPU、内存、GPU...）的分配。它不列出使用情况，如 kubectl top。它可以按命名空间、节点、Pod 分组结果，并通过资源名称进行过滤。

显示的列

• Requested : Pod 清单中容器请求的资源量。它是按 Pod、命名空间、运行容器的节点分组求和。包括组内请求的资源占可分配资源的百分比。
• Limit : Pod 清单中容器最大（限制）请求的资源量。它是按 Pod、命名空间、运行容器的节点分组求和。包括组内最大/限制资源占可分配资源的百分比。
• Allocatable : 节点上定义（或检测到的）可分配资源。
• Free : Allocatable - max (Limit, Requested)
• Utilization : 由 Metrics API 报告的资源（CPU 和内存）使用量。默认禁用，metrics-server 是可选的，应配置在集群中。

安装

通过下载二进制文件

从 GitHub 发布版 下载或使用脚本

curl https://raw.githubusercontent.com/davidB/kubectl-view-allocations/master/scripts/getLatest.sh | bash

通过 krew（kubectl 插件管理器）

Krew – kubectl 插件管理器

kubectl krew install view-allocations

通过 cargo

cargo install kubectl-view-allocations

作为 Cargo.toml 中的库

如果您想将插件的一些函数或结构嵌入到其他 Rust 代码中

[dependencies]
kubectl-view-allocations = { version = "0.14", default-features = false }

[features]
default = ["k8s-openapi/v1_20"]

用法

显示帮助

> kubectl-view-allocations -h
kubectl plugin to list allocations (cpu, memory, gpu,... X utilization, requested, limit, allocatable,...)

Usage: kubectl-view-allocations [OPTIONS]

Options:
      --context <CONTEXT>              The name of the kubeconfig context to use
  -n, --namespace <NAMESPACE>          Show only pods from this namespace
  -u, --utilization                    Force to retrieve utilization (for cpu and memory), require to have metrics-server https://github.com/kubernetes-sigs/metrics-server
  -z, --show-zero                      Show lines with zero requested and zero limit and zero allocatable
      --precheck                       pre-check access and refersh token on kubeconfig by running `kubectl cluster-info`
      --accept-invalid-certs           accept invalid certificats (dangerous)
  -r, --resource-name <RESOURCE_NAME>  Filter resources shown by name(s), by default all resources are listed
  -g, --group-by <GROUP_BY>            Group information hierarchically (default: -g resource -g node -g pod) [possible values: resource, node, pod, namespace]
  -o, --output <OUTPUT>                Output format [default: table] [possible values: table, csv]
  -h, --help                           Print help
  -V, --version                        Print version

https://github.com/davidB/kubectl-view-allocations

显示 GPU 分配

> kubectl-view-allocations -r gpu

 Resource                   Requested       Limit  Allocatable  Free
  nvidia.com/gpu           (71%) 10.0  (71%) 10.0         14.0   4.0
  ├─ node-gpu1               (0%)  __    (0%)  __          2.0   2.0
  ├─ node-gpu2               (0%)  __    (0%)  __          2.0   2.0
  ├─ node-gpu3             (100%) 2.0  (100%) 2.0          2.0    __
  │  └─ fah-gpu-cpu-d29sc         2.0         2.0           __    __
  ├─ node-gpu4             (100%) 2.0  (100%) 2.0          2.0    __
  │  └─ fah-gpu-cpu-hkg59         2.0         2.0           __    __
  ├─ node-gpu5             (100%) 2.0  (100%) 2.0          2.0    __
  │  └─ fah-gpu-cpu-nw9fc         2.0         2.0           __    __
  ├─ node-gpu6             (100%) 2.0  (100%) 2.0          2.0    __
  │  └─ fah-gpu-cpu-gtwsf         2.0         2.0           __    __
  └─ node-gpu7             (100%) 2.0  (100%) 2.0          2.0    __
     └─ fah-gpu-cpu-x7zfb         2.0         2.0           __    __

仅概述

> kubectl-view-allocations -g resource

 Resource              Requested          Limit  Allocatable     Free
  cpu                 (21%) 56.7    (65%) 176.1        272.0     95.9
  ephemeral-storage     (0%)  __       (0%)  __        38.4T    38.4T
  memory             (8%) 52.7Gi  (15%) 101.3Gi      675.6Gi  574.3Gi
  nvidia.com/gpu      (71%) 10.0     (71%) 10.0         14.0      4.0
  pods                (9%) 147.0     (9%) 147.0         1.6k     1.5k

显示利用情况

• 利用率信息从metrics-server（应在您的集群上设置）中检索。
• 仅报告CPU和内存利用率。

> kubectl-view-allocations -u

 Resource                                        Utilization     Requested         Limit  Allocatable   Free 
  cpu                                              (0%) 9.0m  (10%) 200.0m            __          2.0    1.8 
  └─ lima-rancher-desktop                          (0%) 9.0m  (10%) 200.0m            __          2.0    1.8 
     ├─ coredns-96cc4f57d-57cj9                         1.0m        100.0m            __           __     __ 
     ├─ local-path-provisioner-84bb864455-czzcg         1.0m            __            __           __     __ 
     ├─ metrics-server-ff9dbcb6c-kb7x9                  4.0m        100.0m            __           __     __ 
     ├─ svclb-traefik-ggd2q                             2.0m            __            __           __     __ 
     └─ traefik-55fdc6d984-sqp57                        1.0m            __            __           __     __ 
  ephemeral-storage                                       __            __            __        99.8G     __ 
  └─ lima-rancher-desktop                                 __            __            __        99.8G     __ 
  memory                                         (1%) 51.0Mi  (2%) 140.0Mi  (3%) 170.0Mi        5.8Gi  5.6Gi 
  └─ lima-rancher-desktop                        (1%) 51.0Mi  (2%) 140.0Mi  (3%) 170.0Mi        5.8Gi  5.6Gi 
     ├─ coredns-96cc4f57d-57cj9                       11.5Mi        70.0Mi       170.0Mi           __     __ 
     ├─ local-path-provisioner-84bb864455-czzcg        6.2Mi            __            __           __     __ 
     ├─ metrics-server-ff9dbcb6c-kb7x9                14.9Mi        70.0Mi            __           __     __ 
     ├─ svclb-traefik-ggd2q                          548.0Ki            __            __           __     __ 
     └─ traefik-55fdc6d984-sqp57                      17.9Mi            __            __           __     __ 
  pods                                                    __      (5%) 5.0      (5%) 5.0        110.0  105.0 
  └─ lima-rancher-desktop                                 __      (5%) 5.0      (5%) 5.0        110.0  105.0

按命名空间分组。

> kubectl-view-allocations -g namespace

 Resource               Requested         Limit  Allocatable   Free 
  cpu                (10%) 200.0m            __          2.0    1.8 
  └─ kube-system           200.0m            __           __     __ 
  ephemeral-storage            __            __        99.8G     __ 
  memory             (2%) 140.0Mi  (3%) 170.0Mi        5.8Gi  5.6Gi 
  └─ kube-system          140.0Mi       170.0Mi           __     __ 
  pods                   (5%) 5.0      (5%) 5.0        110.0  105.0 
  └─ kube-system              5.0           5.0           __     __ 

以CSV格式显示。

在这种情况下，值以浮点数形式展开（保留2位小数）。

kubectl-view-allocations -o csv
Date,Kind,resource,node,pod,Requested,%Requested,Limit,%Limit,Allocatable,Free
2020-08-19T19:12:48.326605746+00:00,resource,cpu,,,59.94,22%,106.10,39%,272.00,165.90
2020-08-19T19:12:48.326605746+00:00,node,cpu,node-gpu1,,2.31,19%,4.47,37%,12.00,7.53
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu1,yyy-b8bd56fbd-5x8vq,1.00,,2.00,,,
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu1,kube-flannel-ds-amd64-7dz9z,0.10,,0.10,,,
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu1,node-exporter-gpu-b4w7s,0.11,,0.22,,,
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu1,xxx-backend-7d84544458-46qnh,1.00,,2.00,,,
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu1,weave-scope-agent-bbdnz,0.10,,0.15,,,
2020-08-19T19:12:48.326605746+00:00,node,cpu,node-gpu2,,0.31,1%,0.47,2%,24.00,23.53
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu2,kube-flannel-ds-amd64-b5b4v,0.10,,0.10,,,
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu2,node-exporter-gpu-796jz,0.11,,0.22,,,
2020-08-19T19:12:48.326605746+00:00,pod,cpu,node-gpu2,weave-scope-agent-8rhnd,0.10,,0.15,,,
2020-08-19T19:12:48.326605746+00:00,node,cpu,node-gpu3,,3.41,11%,6.67,21%,32.00,25.33
...

它可以与“分组”选项结合使用。

kubectl-view-allocations -g resource -o csv
Date,Kind,resource,Requested,%Requested,Limit,%Limit,Allocatable,Free
2020-08-19T19:11:49.630864028+00:00,resource,cpu,59.94,22%,106.10,39%,272.00,165.90
2020-08-19T19:11:49.630864028+00:00,resource,ephemeral-storage,0.00,0%,0.00,0%,34462898618662.00,34462898618662.00
2020-08-19T19:11:49.630864028+00:00,resource,hugepages-1Gi,0.00,,0.00,,,
2020-08-19T19:11:49.630864028+00:00,resource,hugepages-2Mi,0.00,,0.00,,,
2020-08-19T19:11:49.630864028+00:00,resource,memory,69063409664.00,10%,224684670976.00,31%,722318667776.00,497633996800.00
2020-08-19T19:11:49.630864028+00:00,resource,nvidia.com/gpu,3.00,27%,3.00,27%,11.00,8.00
2020-08-19T19:11:49.630864028+00:00,resource,pods,0.00,0%,0.00,0%,1540.00,1540.00

替代方案及类似产品

• 查看讨论需要简单的kubectl命令来查看集群资源使用情况 · Issue #17512 · kubernetes/kubernetes
• 仅针对CPU和内存
  • robscott/kube-capacity：一个简单的CLI，它提供Kubernetes集群中资源请求、限制和利用率的概述,
  • hjacobs/kube-resource-report：报告Kubernetes集群和Pod的资源请求与使用情况，并生成静态HTML
  • etopeter/kubectl-view-utilization：kubectl插件，用于显示集群CPU和内存请求利用率
• 仅针对CPU和内存利用率
  • kubectl top pods
  • LeastAuthority/kubetop：Kubernetes的类似top(1)的工具。