kubectl plugin lists allocations for resources (cpu, memory, gpu,...) as defined into the manifest of nodes and running pods. It doesn't list usage like kubectl top. It can provide result grouped by namespaces, nodes, pods and filtered by resources'name.
Columns displayed :
Requested: Quantity of resources requested by the container in the pod's manifest. It's the sum group by pod, namespace, node where container is running. With percentage of resources requested over what is allocatable in the group.Limit: Quantity of resources max (limit) requestable by the container in the pod's manifest. It's the sum group by pod, namespace, node where container is running. With percentage of resources max / limit over what is allocatable in the group.Allocatable: Allocatable resources defined (or detected) on nodes.Free:Allocatable - max (Limit, Requested)(by default, see options--used-mode)Utilization: Quantity of resources (cpu & memory only) used as reported by Metrics API. It's disable by default, metrics-server is optional and should be setup into the cluster.
Download from github's release or use script
curl https://raw.githubusercontent.com/davidB/kubectl-view-allocations/master/scripts/getLatest.sh | bashkubectl krew install view-allocationscargo install kubectl-view-allocationsIf you want to embed some function or struct of the plugin into an other rust code:
[dependencies]
kubectl-view-allocations = { version = "0.14", default-features = false }
[features]
default = ["k8s-openapi/v1_20"]> kubectl-view-allocations -h
kubectl plugin to list allocations (cpu, memory, gpu,...) X (utilization, requested, limit, allocatable,...)
Usage: kubectl-view-allocations [OPTIONS]
Options:
--kubeconfig <KUBECONFIG>
Path to the kubeconfig file to use for requests to kubernetes cluster
--context <CONTEXT>
The name of the kubeconfig context to use
-n, --namespace <NAMESPACE>...
Filter pods by namespace(s), by default pods in all namespaces are listed (comma separated list or multiple calls)
-l, --selector <SELECTOR>
Show only nodes match this label selector
-u, --utilization
Force to retrieve utilization (for cpu and memory), requires having metrics-server https://github.com/kubernetes-sigs/metrics-server
-z, --show-zero
Show lines with zero requested AND zero limit AND zero allocatable, OR pods with unset requested AND limit for `cpu` and `memory`
--used-mode <USED_MODE>
The way to compute the `used` part for free (`allocatable - used`) [default: max-request-limit] [possible values: max-request-limit, only-request]
--precheck
Pre-check access and refresh token on kubeconfig by running `kubectl cluster-info`
--accept-invalid-certs
Accept invalid certificates (dangerous)
-r, --resource-name <RESOURCE_NAME>...
Filter resources shown by name(s), by default all resources are listed (comma separated list or multiple calls)
-g, --group-by <GROUP_BY>...
Group information in a hierarchical manner; defaults to `-g resource,node,pod` (comma-separated list or multiple calls) [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> 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 __ __> 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- Utilization information are retrieve from metrics-server (should be setup on your cluster).
- Only report cpu and memory utilization
> 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> 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 __ __In this case value as expanded as float (with 2 decimal)
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
...It can be combined with "group-by" options.
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- see the discussion Need simple kubectl command to see cluster resource usage · Issue #17512 · kubernetes/kubernetes
- For CPU & Memory only
- ahmetb/kubectl-node_resource: Query node allocations/utilization in kubectl
- robscott/kube-capacity: A simple CLI that provides an overview of the resource requests, limits, and utilization in a Kubernetes cluster,
- hjacobs/kube-resource-report: Report Kubernetes cluster and pod resource requests vs usage and generate static HTML
- etopeter/kubectl-view-utilization: kubectl plugin to show cluster CPU and Memory requests utilization
- For CPU & Memory utilization only
kubectl top pods- LeastAuthority/kubetop: A top(1)-like tool for Kubernetes.