Kram - Kubernetes Resource Metrics Printer

Kram is a command-line tool that retrieves resource metrics for Kubernetes namespaces and pods and prints them in a tabular format.

Last Build

Prerequisites

Before using this application, ensure you have the following prerequisites:

• Go installed on your system. (to build)
• kubectl configured with access to your Kubernetes cluster.
• metrics-server deployed in your cluster.

Installation

1. Clone the repository to your local machine:

git clone https://github.com/PaulPowershell/Kram
cd Kram

2. Build the Go application:

go build .

Download Kram Executable

You can download the executable for Kram directly from the latest release with its version. This allows you to use Kram without the need to build it yourself. Here are the steps to download the executable for your system:

1. Visit the Releases page.

Usages

commands:

Usage:
  kram [namespace] [flags]

Flags:
  -c, --cpu                 Show only CPU table (use with -N)
  -h, --help                help for kram
  --kubeconfig string   (optional) string Absolute path to the kubeconfig file (default "~/.kube/config")
  -N, --node                Display resource usage matrix by node
  -o, --output string       Output format: table or html (default "table")
  -r, --ram                 Show only RAM table (use with -N)

Example 1: List metrics for all namespaces

To list metrics for all namespaces, run the application without any arguments:

kram

The application outputs (kram) the following metrics in a tabular format:

Namespace	Pods	CPU Usage	CPU Request	CPU Limit	Mem Usage	Mem Request	Mem Limit
example-namespace	5	100 m	200 m	300 m	500 MiB	600 MiB	700 MiB
another-namespace	3	50 m	100 m	150 m	250 MiB	300 MiB	350 MiB
Total	8	150 m	300 m	350 m	750 MiB	900 MiB	1.05 GiB

Example 2: List metrics for a specific namespace

To list metrics for a specific namespace, provide the namespace name as an argument:

kram <namespace>

Metrics for Namespace: networking

Pods	Container	CPU Usage	CPU Request	CPU Limit	Mem Usage	Mem Request	Mem Limit
ingress-nginx-controller-xxxxxxxxxx-xxxxx	controller	2 m	100 m	100 m	62.09MiB	256MiB	512MiB
ingress-nginx-controller-xxxxxxxxxx-xxxxx	controller	3 m	100 m	100 m	62.71MiB	256MiB	512MiB
ingress-nginx-defaultbackend-xxxxxxxxxxxx-xxx	ingress-nginx-default-backend	1 m	0 m	0 m	4.734MiB	0B	0B
Total		6 m	200 m	200 m	129.5MiB	512MiB	1GiB

Example 3: List metrics by namespaces on nodes

To list metrics by namespaces on nodes:

kram --node

Memory Usage / Request / Limit

Namespace	aks-computespot-xxxxxxxx-xxxxxxxxxx	aks-computespot-xxxxxxxx-xxxxxxxxxx	aks-sys-xxxxxxxx-xxxxxxxxxx
flux-system	-	-	679.3MiB/400MiB/6.016GiB
kube-system	302.1MiB/446MiB/7.482GiB	245.9MiB/446MiB/7.482GiB	518.3MiB/970MiB/13.26GiB
monitoring	256.4MiB/322MiB/1.064GiB	1.182GiB/1.549GiB/2.799GiB	172MiB/336MiB/1.123GiB
networking	4.734MiB/0B/0B	-	125.4MiB/512MiB/1GiB
opencost	108.6MiB/71MiB/272MiB	-	-

Example 4: List metrics for a specific namespace by nodes

To list metrics for a specific namespace by nodes, provide the namespace name as an argument:

kram <namespace> --node

Memory Usage / Request / Limit

Namespace	aks-computespot-xxxxxxxx-xxxxxxxxxx	aks-sys-xxxxxxxx-xxxxxxxxxx
networking	4.734MiB/0B/0B	123.3MiB/512MiB/1GiB

CPU Usage / Request / Limit

Namespace	aks-computespot-xxxxxxxx-xxxxxxxxxx	aks-sys-xxxxxxxx-xxxxxxxxxx
networking	1m/0m/0m	5m/200m/200m

Example 5: List ram or metrics for a specific namespace by nodes

To list ram (or cpu) metrics for a specific namespace by nodes, provide the namespace name as an argument:

kram <namespace> --node --ram

Memory Usage / Request / Limit

Namespace	aks-computespot-xxxxxxxx-xxxxxxxxxx	aks-sys-xxxxxxxx-xxxxxxxxxx
networking	4.734MiB/0B/0B	123.3MiB/512MiB/1GiB

Example 6: Export metrics to HTML

Any command can be combined with --output html (or -o html) to generate an HTML report and automatically open it in the default browser.

# Node view → kram.html
kram --node --output html

The HTML report uses a dark theme and renders the same tables in a responsive, browser-friendly format.

License

This project is licensed under the MIT License. See the LICENSE file for details.