kubetui

Kubetui is a terminal user interface (TUI) tool designed for monitoring Kubernetes resources.
It provides an easy-to-use interface for developers and operators to access important information about their applications and infrastructure.

Features

Kubetui offers the following features to help you monitor and manage your Kubernetes resources:

• Pods List and Container Logs: Browse pods and stream their container logs, with JSON pretty-print toggling (f/p) and a powerful Log Query (regex, label/field selectors, resource targeting, jq/JMESPath).
• Node List and Detail: View nodes with status, roles, age, and version, plus a detail pane.
• ConfigMap and Secret Watching: Monitor ConfigMaps and Secrets, and decode their (Base64) data.
• Network-related Resources: List Services, Ingresses, NetworkPolicies, and Gateway API resources (Gateway / HTTPRoute), and view details for the selected one.
• Events Watching: Stay updated with a real-time view of Kubernetes events.
• Arbitrary Resource Watching (List / YAML): Select any resource kinds with f and watch them as a list, or inspect a selected resource's raw YAML with y.
• Customizable Columns: Pick visible columns and order via a runtime dialog (t), CLI flags / presets (Pod / Node), and label columns. See Column Customization.
• Column-aware Filter: Filter table rows by column with include/exclude regex and server-side label selectors, with inline help. See Filter.
• Namespace Multiple Selections: Select and view multiple namespaces simultaneously.
• Context Selection: Switch the Kubernetes context you operate on (with namespace carry-over / caching).
• Adjustable Split Layout: Toggle vertical/horizontal pane split at runtime (Shift+s) or at startup (-s v|h).
• Clipboard Support: Copy text with the mouse; the backend is selectable (system / OSC52, SSH- and tmux-friendly). See Clipboard.
• Mouse Support: Click to focus and select, click tabs to switch, scroll with the wheel, and drag to select text for copying.
• Incremental Search: Search within text views with / and jump between matches with n / N.
• (beta) Customizable UI Appearance: Theme border styles, colors, and text attributes via a config file.

Overall, kubetui is a powerful tool designed to provide a safe and efficient way to access and monitor your Kubernetes resources. With its user-friendly interface and comprehensive features, it simplifies the process of managing your applications and infrastructure.

Installation

To install kubetui, you can use the following methods:

Homebrew on macOS and Linux

Kubetui is available on homebrew, the package manager for macOS and Linux. Install it by running the following command:

brew install kubetui

Pacman on Arch Linux

Kubetui is available in the official repositories. Install it by running the following command:

pacman -S kubetui

Scoop on Windows

If you are using Windows with scoop, you can add the necessary buckets and install kubetui with the following commands:

# Add the 'extras' bucket for vcredist2022
scoop bucket add extras
scoop bucket add <bucket> https://github.com/sarub0b0/scoop-bucket
scoop install <bucket>/kubetui

WinGet on Windows

If you prefer using winget, the Windows package manager, you can install kubetui with the following command:

winget install kubetui

Chocolatey on Windows

Kubetui is available on Chocolatey, the package manager for Windows. Install it by running the following command:

choco install kubetui

openSUSE Tumbleweed

For openSUSE Tumbleweed, you can install kubetui using the zypper package manager. Run the following command to install:

zypper install kubetui

Using cargo install

Kubetui is available on crates.io, the official Rust package registry. Install it by running the following command:

cargo install kubetui

Make sure you have Rust and Cargo installed before running the command.

Downloading the binary

Alternatively, you can download the precompiled binary from the GitHub Release page that matches your operating system. Once downloaded, you can run the binary directly without any additional installation steps.

Choose the method that suits your needs and preferences.

Usage

kubetui

$ kubetui -h
An intuitive Terminal User Interface (TUI) tool for real-time monitoring and exploration of Kubernetes resources.

Usage: kubetui [OPTIONS] [COMMAND]

Commands:
  completion  Generate completion script

Options:
  -h, --help                                       Print help
  -V, --version                                    Print version
  -A, --all-namespaces[=<true|false>]              Select all namespaces [default: false]
  -c, --context <CONTEXT>                          Context
  -C, --kubeconfig <KUBECONFIG>                    kubeconfig path
      --clipboard <auto|system|osc52>              Clipboard mode (auto, system, or osc52) [env: KUBETUI_CLIPBOARD=] [default: auto]
      --config-file <CONFIG_FILE>                  Config file path
  -l, --logging                                    Logging
  -n, --namespaces <NAMESPACES>                    Namespaces (e.g. -n val1,val2,val3 | -n val1 -n val2 -n val3)
      --node-columns <NODE_COLUMNS>                Comma-separated columns for the node table: builtin names (e.g. name,status), defined label-column names, or "full" for all builtins
      --node-columns-preset <NODE_COLUMNS_PRESET>  Preset name for node columns (e.g. "default", "wide"). If both are specified, `--node-columns` overrides this
      --pod-columns <POD_COLUMNS>                  Comma-separated list of columns to show in pod table (e.g. name,status,ip). Use "full" to show all available columns
      --pod-columns-preset <POD_COLUMNS_PRESET>    Preset name for pod columns (e.g. "default", "full"). If both are specified, `--pod-columns` overrides this
  -s, --split-direction <v|h>                      Window split direction [default: v]

Column Customization

The Pod, Node, Config, and Network tables all support column customization. Pod and Node additionally support presets and a CLI flag for selecting columns at startup; Config and Network use the runtime dialog and label columns.

CLI flags (Pod / Node)

kubetui --pod-columns=name,ready,status,age
kubetui --pod-columns=full           # show all builtin columns
kubetui --pod-columns-preset=default

kubetui --node-columns=name,status,roles,age,version
kubetui --node-columns-preset=wide

Notes:

• The NAME column is always included even if not specified.
• full expands to all builtin columns and cannot be combined with other columns.
• --pod-columns overrides --pod-columns-preset; same for --node-columns / --node-columns-preset.

Runtime column dialog (all four tabs)

Press t while a table is focused (Pod / Node / Config / Network) to open the column selection dialog.

• Space or Enter: toggle visibility
• J / K: reorder columns
• Required columns like NAME are always enabled and fixed.

Label columns

You can register labels as table columns under theme.<tab>.label_columns. Each entry maps a short name to a label key. The name is used as the upper-cased column header, in the column dialog, and in filter expressions across all four tabs; for Pod and Node it can additionally appear in --*-columns flag values and presets. The cell value is taken from metadata.labels[<label>]; resources without the label show an empty cell.

theme:
  pod:
    label_columns:
      - name: app
        label: app.kubernetes.io/name

  node:
    label_columns:
      - name: zone
        label: topology.kubernetes.io/zone
      - name: instance
        label: node.kubernetes.io/instance-type

  config:
    label_columns:
      - name: instance
        label: argocd.argoproj.io/instance

  network:
    label_columns:
      - name: app
        label: app.kubernetes.io/name

Registered label columns are also valid:

• Inside --pod-columns / --node-columns (e.g. --node-columns=name,status,zone).
• Inside presets (interleaved with builtin columns).
• Inside filter expressions (e.g. app:nginx, zone:asia-northeast1-a).

Define presets in config.yaml (Pod / Node)

Define reusable presets under theme.pod.column_presets / theme.node.column_presets, and set a startup default with default_preset.

theme:
  pod:
    default_preset: minimal
    column_presets:
      default:
        - name
        - status
        - age
        - ip
        - node
      minimal:
        - name
        - status
        - age

  node:
    default_preset: default
    column_presets:
      default:
        - name
        - status
        - roles
        - age
        - version
      wide:
        - full
      topology:
        - name
        - status
        - roles
        - zone       # label column
        - instance   # label column

Resolution order at startup: CLI flag (--<tab>-columns) > preset (--<tab>-columns-preset) > theme.<tab>.default_preset > builtin default.

Filter (Column-Aware)

Pod, Node, Config, and Network tables share a column-aware filter. Open the filter input with /, type the expression, and press Enter to apply (or Esc to clear the active filter and close the form).

Press ? (or type help) inside the filter input to open the per-tab filter help dialog with the columns available in the current tab.

Syntax summary

TERM [ TERM ]...

• Same column, multiple includes → OR (in-list): STATUS:Running STATUS:Pending → STATUS in (Running, Pending).
• Different columns, includes → AND across columns: NAME:web STATUS:Running → NAME~web AND STATUS~Running.
• Any matching exclude → row excluded.
• Column names ignore case, spaces, -, and _.
• Values with whitespace must be quoted: STATUS:"CreateContainerConfigError". Escape ", ', or \ inside quotes with \.

Inactive terms

A term on a column that is currently not shown stays in the filter but is not applied. The title shows (inactive: COL) until the column is shown again (e.g. via the column dialog). Unknown columns produce an error.

Notes per tab

• Pod / Config / Network: namespace is not filterable — use the namespace selector (n / N) instead. A namespace:<...> term returns a dedicated guidance message (rather than a generic unknown-column error) in all three tabs.
• Node: cluster-scoped, so there is no namespace concept; namespace:<...> is treated as a plain unknown-column error.

Shell Completion

Kubetui supports shell completion for Bash and Zsh. You can enable the completion by adding the following to your shell configuration file:

For Bash (add to ~/.bashrc or ~/.bash_profile):

source <(kubetui completion bash)

For Zsh (add to ~/.zshrc):

source <(kubetui completion zsh)

Clipboard

Kubetui supports three clipboard backends, selectable via --clipboard or the KUBETUI_CLIPBOARD environment variable:

kubetui --clipboard osc52
KUBETUI_CLIPBOARD=osc52 kubetui

Custom Configuration

You can customize the UI appearance and several feature settings by specifying a configuration file using the --config-file flag:

kubetui --config-file /path/to/your/config.yaml

The configuration file can also be located at $XDG_CONFIG_HOME/kubetui/config.yaml (or ~/.config/kubetui/config.yaml when $XDG_CONFIG_HOME is unset).

The configuration file allows you to modify:

• Border Styles: Customize the border styles of different UI components.
• Colors: Change the colors of text, backgrounds, and borders.
• Text Attributes: Modify text attributes such as bold, italic, and underline.
• Per-tab settings: theme.pod / theme.node / theme.config / theme.network accept label_columns (register labels as columns and filter terms) and column_presets / default_preset (Pod and Node only).
• Status highlights: theme.pod.highlights and theme.event.highlights accept regex → style rules.

A sample configuration file is available at example/config.yaml to help you get started.

Log Query

The Log Query feature empowers you to retrieve logs from multiple Pods and their containers. Using regular expressions, selectors, and specified resources, you can precisely define the log retrieval targets. This functionality also allows you to filter logs using regular expressions, providing a powerful and flexible log querying experience.

Usage Example

pod:app container:nginx log:401

pod:api log:error jq:.message

pod:api log:error jmespath:message

When entering ? or help in the log query form, the help dialog will be displayed.

Supported Queries

Supported resources:

Query String Escaping

When including spaces in queries such as <regex> or <selector>, enclose the string with " or '. For example:

pod:"a b"
label:"environment in (production, qa)"

If you use ", ', or \ within the quoted string, escape them with \. For example:

pod:"a\\b"

**Lexer and Parser**

LOG_QUERIES = QUERY ( " "+ QUERY )*

QUERY = POD
        | EXCLUDE_POD
        | CONTAINER
        | EXCLUDE_CONTAINER
        | LOG
        | EXCLUDE_LOG
        | LABEL
        | FIELD
        | JQ
        | JMESPATH
        | LIMIT
        | SPECIFIED_RESOURCE

POD = ( "pods" | "pod" | "po" | "p" ) ":" REGEX
EXCLUDE_POD = "!" POD

CONTAINER = ( "containers" | "container" | "co" | "c" ) ":" REGEX
EXCLUDE_CONTAINER = "!" CONTAINER

LOG = ( "logs" | "log" | "lo" | "l" ) ":" REGEX
EXCLUDE_LOG = "!" LOG

REGEX = QUOTED_STRING | UNQUOTED_STRING

LABEL = ( "labels" | "label" ) ":" SELECTOR
FIELD = ( "fields" | "field" ) ":" SELECTOR

SELECTOR = QUOTED_STRING | UNQUOTED_STRING

JQ = "jq" ":" EXPR

JMESPATH = ( "jmespath" | "jmes" | "jm" ) ":" EXPR

EXPR = QUOTED_STRING | UNQUOTED_STRING

LIMIT = ( "limit" | "lim" ) ":" POSITIVE_INTEGER

POSITIVE_INTEGER = [1-9] [0-9]*

SPECIFIED_RESOURCE = RESOURCE "/" NAME

RESOURCE = ( "pods" | "pod" | "po" )
           | ( "replicasets" | "replicaset" | "rs" )
           | ( "deployments" | "deployment" | "deploy" )
           | ( "statefulsets" | "statefulset" | "sts" )
           | ( "daemonsets" | "daemonset" | "ds" )
           | ( "services" | "service" | "svc" )
           | ( "jobs" | "job" )

NAME = ALPHANUMERIC ( ALPHANUMERIC | "-" | "." )* ALPHANUMERIC

UNQUOTED_STRING = ~['" \t\r\n] ( ~[ \t\r\n] )* // without spaces

QUOTED_STRING = "\"" ESCAPED_STRING "\"" | "'" ESCAPED_STRING "'"

ESCAPED_STRING = ( ESCAPED_CHAR | ~[\"'] )\*

ESCAPED_CHAR = "\\" | "\"" | "\'"

Key Bindings

General

Key Map

View Control

Text View

Search Mode

Table View

Column Dialog

Dialog

Context Dialog

Input Form

Container Logs View

Inline notices

Two kinds of inline [kubetui] lines may appear within the log stream:

• Yellow [kubetui] <namespace>: <message>: Per-namespace non-fatal notice during setup (e.g. the resource specified by deployment/<name> does not exist in some of the selected namespaces). Other namespaces continue to stream logs.
• Red [kubetui] <message>: A stream-side error encountered while logs are flowing. The stream continues; the error state of the widget is not toggled.

Contributing

Bug reports and pull requests are welcome.

License

This software is available as open source under the terms of the MIT License.