No description
Find a file
Clement Tsang fd003f84da
bug: Fix some performance regressions (#344)
Fixes some performance regressions and forgotten cleanup.

Changes to attempt to improve performance to match 0.4.x:

- Remove `trace!` and `--debug` for now.  These were a significant hog.  Removing this dropped initial memory usage by about half.
- Add additional cleaning step for `pid_mapping`  during process harvesting.  This should hopefully improve memory usage as time goes on.
- Slightly change how we do sorting to hopefully be a bit more optimal?  This was just an easy change to make that I spotted.
- Fix broken cleaning child thread task.
2020-12-10 22:29:25 -05:00
.cargo ci: uptick to 0.5.1, fix ci 2020-11-22 02:33:48 -05:00
.cargo-husky/hooks refactor: Clean up some parts of data harvesting (#336) 2020-12-03 23:45:54 -05:00
.github ci: Switch to musl for homebrew linux (#347) 2020-12-10 20:11:34 -05:00
.vscode refactor: Simplify data harvesting (#335) 2020-11-30 19:40:17 -05:00
assets other: update gruvbox and some documentation (#310) 2020-11-19 23:32:57 -05:00
deployment ci: Switch to musl for homebrew linux (#347) 2020-12-10 20:11:34 -05:00
sample_configs other: update default config.toml 2020-11-22 13:26:58 -05:00
src bug: Fix some performance regressions (#344) 2020-12-10 22:29:25 -05:00
tests ci: uptick to 0.5.1, fix ci 2020-11-22 02:33:48 -05:00
.all-contributorsrc docs: add dm9pZCAq as a contributor (#301) 2020-11-14 17:37:15 -05:00
.gitignore bug: Fix some performance regressions (#344) 2020-12-10 22:29:25 -05:00
.markdownlint.json docs: Add changelog 2020-04-05 23:41:05 -04:00
build.rs other: Add autocomplete file generation (#213) 2020-08-31 17:50:21 -04:00
Cargo.lock refactor: Clean up some parts of data harvesting (#336) 2020-12-03 23:45:54 -05:00
Cargo.toml bug: Fix some performance regressions (#344) 2020-12-10 22:29:25 -05:00
CHANGELOG.md bug: Fix some performance regressions (#344) 2020-12-10 22:29:25 -05:00
clippy.toml refactor: rewrite column algorithm (#227) 2020-09-09 21:51:52 -04:00
CONTRIBUTING.md feature: Beginnings of in-app config (#231) 2020-09-22 18:12:36 -04:00
FAQ.md docs: Add FAQ document (#339) 2020-12-04 20:11:38 -05:00
LICENSE Initial commit 2019-08-28 19:43:30 -04:00
README.md bug: Fix some performance regressions (#344) 2020-12-10 22:29:25 -05:00
rustfmt.toml ci: uptick to 0.5.1, fix ci 2020-11-22 02:33:48 -05:00

bottom

ci crates.io link

A cross-platform graphical process/system monitor with a customizable interface and a multitude of features. Supports Linux, macOS, and Windows. Inspired by both gtop and gotop.

Quick demo recording showing off searching, expanding, and process killing. Theme based on gruvbox (--color gruvbox). Font is IBM Plex Mono, terminal is Kitty.

Note: If you are reading this on the master branch, then it may refer to in-development or un-released features/changes. Please refer to release branch or crates.io for the most up-to-date release documentation.

Table of Contents

Installation

Note that bottom is:

  • Built on the stable version of Rust
  • Officially tested and released for only x86_64 (and i686 for Windows)
  • Developed mainly for macOS, Windows, and Linux

Anything outside of this (i.e: ARM builds, building on Nightly, building on another OS) is currently not guaranteed, even if it does happen to work. For example, ARM is compiled on the CI pipeline and release builds will be provided, but not all features may work (such as R/s and W/s for disks).

Manually

There are a few ways to go about doing this manually. If you do so, please build using the current stable release of Rust. For example:

# If required, update Rust on the stable channel
rustup update stable

# Clone and install the newest master version all via Cargo
cargo install --git https://github.com/ClementTsang/bottom

# Clone from master and install manually
git clone https://github.com/ClementTsang/bottom
cd bottom
cargo install --path .

# Download from releases and install
curl -LO https://github.com/ClementTsang/bottom/archive/0.5.3.tar.gz
tar -xzvf 0.5.3.tar.gz
cargo install --path .

Or, you can just download the binary from the latest release.

Cargo

# If required, update Rust on the stable channel
rustup update stable

cargo install bottom

# OR, --locked may be required due to how cargo install works
cargo install bottom --locked

AUR

yay -S bottom

# If you instead want a pre-built binary:
yay -S bottom-bin

Debian

A .deb file is provided on each release:

curl -LO https://github.com/ClementTsang/bottom/releases/download/0.5.3/bottom_0.5.3_amd64.deb
sudo dpkg -i bottom_0.5.3_amd64.deb

Fedora/CentOS

Available in COPR:

sudo dnf copr enable atim/bottom -y
sudo dnf install bottom

Gentoo

Available in dm9pZCAq overlay

sudo eselect repository enable dm9pZCAq
sudo emerge --sync dm9pZCAq
sudo emerge sys-process/bottom::dm9pZCAq

Nix

nix-env -i bottom

Homebrew

brew tap clementtsang/bottom
brew install bottom

# If you need to be more specific, use:
brew install clementtsang/bottom/bottom

Scoop

scoop install bottom

Chocolatey

Choco package located here. Since you cannot upload a new package while a previous one is being validated, the newest version of a release may take a while to appear.

choco install bottom

# Version number may be required for newer releases, if available:
choco install bottom --version=0.5.3

winget

You can find the packages here. Since validation of the package takes time, it may take a while to become available after a release.

winget install bottom

You can also manually do the same thing by going to the latest release and installing via the .msi file.

You can uninstall via Control Panel or Options in Windows.

Auto-completion

Shell completions are included in binary releases, and are generated in the same directory as the binary if bottom is manually built.

  • For bash, move btm.bash to $XDG_CONFIG_HOME/bash_completion or /etc/bash_completion.d/.
  • For fish, move btm.fish to $HOME/.config/fish/completions/.
  • For zsh, move _btm to one of your $fpath directories.
  • For PowerShell, add . _btm.ps1 to your PowerShell profile.

Some install scripts (i.e. AUR) will automatically do this for you.

Usage

Run using btm.

Flags

Use btm --help for more information.

        --autohide_time                        Temporarily shows the time scale in graphs.
    -b, --basic                                Hides graphs and uses a more basic look.
        --battery                              Shows the battery widget.
    -S, --case_sensitive                       Enables case sensitivity by default.
    -c, --celsius                              Sets the temperature type to Celsius.
        --color <COLOR SCHEME>                 Use a color scheme, use --help for supported values.
    -C, --config <CONFIG PATH>                 Sets the location of the config file.
    -u, --current_usage                        Sets process CPU% to be based on current CPU%.
    -t, --default_time_value <MS>              Default time value for graphs in ms.
        --default_widget_count <INT>           Sets the n'th selected widget type as the default.
        --default_widget_type <WIDGET TYPE>    Sets which widget type to use as the default widget.
        --disable_click                        Disables mouse clicks.
    -m, --dot_marker                           Uses a dot marker for graphs.
    -f, --fahrenheit                           Sets the temperature type to Fahrenheit.
    -g, --group                                Groups processes with the same name by default.
    -h, --help                                 Prints help information.  Use --help for more info.
    -a, --hide_avg_cpu                         Hides the average CPU usage.
        --hide_table_gap                       Hides the spacing between table headers and entries.
        --hide_time                            Completely hides the time scaling.
    -k, --kelvin                               Sets the temperature type to Kelvin.
    -l, --left_legend                          Puts the CPU chart legend to the left side.
        --mem_as_value                         Defaults to showing process memory usage by value.
    -r, --rate <MS>                            Sets a refresh rate in ms.
    -R, --regex                                Enables regex by default.
    -d, --time_delta <MS>                      The amount in ms changed upon zooming.
    -T, --tree                                 Defaults to showing the process widget in tree mode.
        --use_old_network_legend               DEPRECATED - uses the older network legend.
    -V, --version                              Prints version information.
    -W, --whole_word                           Enables whole-word matching by default.

Keybindings

General

q, Ctrl-c Quit
Esc Close dialog windows, search, widgets, or exit expanded mode
Ctrl-r Reset display and any collected data
f Freeze/unfreeze updating with new data
Ctrl-Left
Shift-Left
H
A
Move widget selection left
Ctrl-Right
Shift-Right
L
D
Move widget selection right
Ctrl-Up
Shift-Up
K
W
Move widget selection up
Ctrl-Down
Shift-Down
J
S
Move widget selection down
Left, h Move left within widget
Down, j Move down within widget
Up,k Move up within widget
Right, l Move right within widget
? Open help menu
gg, Home Jump to the first entry
Shift-g, End Jump to the last entry
e Toggle expanding the currently selected widget
+ Zoom in on chart (decrease time range)
- Zoom out on chart (increase time range)
= Reset zoom

Process bindings

dd Kill the selected process
c Sort by CPU usage, press again to reverse sorting order
m Sort by memory usage, press again to reverse sorting order
p Sort by PID name, press again to reverse sorting order
n Sort by process name, press again to reverse sorting order
Tab Group/un-group processes with the same name
Ctrl-f, / Open process search widget
P Toggle between showing the full command or just the process name
s, F6 Open process sort widget
I Invert current sort
% Toggle between values and percentages for memory usage
t, F5 Toggle tree mode

Process search bindings

Tab Toggle between searching by PID or name
Esc Close the search widget (retains the filter)
Ctrl-a Skip to the start of the search query
Ctrl-e Skip to the end of the search query
Ctrl-u Clear the current search query
Backspace Delete the character behind the cursor
Delete Delete the character at the cursor
Alt-c, F1 Toggle matching case
Alt-w, F2 Toggle matching the entire word
Alt-r, F3 Toggle using regex
Left Move cursor left
Right Move cursor right

Process sort bindings

Down, j Scroll down in list
Up, k Scroll up in list
Mouse scroll Scroll through sort widget
Esc Close the sort widget
Enter Sort by current selected column

Battery bindings

Left, Alt-h Go to the next battery
Right, Alt-l Go to the previous battery

Basic memory bindings

% Toggle between values and percentages for memory usage

Process searching keywords

  • None of the keywords are case sensitive.
  • Use brackets to logically group together parts of the search.
  • Furthermore, if you want to search a reserved keyword, surround the text in quotes - for example, "or" or "(sd-pam)" would be a valid search:

quote searching

Supported search types

Keywords Example Description
btm Matches by process or command name; supports regex
pid pid=1044 Matches by PID; supports regex
cpu, cpu% cpu > 0.5 Matches the CPU column; supports comparison operators
memb memb > 1000 b Matches the memory column in terms of bytes; supports comparison operators
mem, mem% mem < 0.5 Matches the memory column in terms of percent; supports comparison operators
read, r/s read = 1 mb Matches the read/s column in terms of bytes; supports comparison operators
write, w/s write >= 1 kb Matches the write/s column in terms of bytes; supports comparison operators
tread, t.read tread <= 1024 gb Matches he total read column in terms of bytes; supports comparison operators
twrite, t.write twrite > 1024 tb Matches the total write column in terms of bytes; supports comparison operators
state state=running Matches by state; supports regex

Supported comparison operators

Keywords Description
= Checks if the values are equal
> Checks if the left value is strictly greater than the right
< Checks if the left value is strictly less than the right
>= Checks if the left value is greater than or equal to the right
<= Checks if the left value is less than or equal to the right

Supported logical operators

Note that the and operator takes precedence over the or operator.

Keywords Usage Description
and, &&, <Space> <CONDITION 1> and/&&/<Space> <CONDITION 2> Requires both conditions to be true to match
or, || <CONDITION 1> or/|| <CONDITION 2> Requires at least one condition to be true to match

Supported units

Keywords Description
B Bytes
KB Kilobytes
MB Megabytes
GB Gigabytes
TB Terabytes
KiB Kibibytes
MiB Mebibytes
GiB Gibibytes
TiB Tebibytes

Other syntax

Keywords Usage Description
() (<CONDITION 1> AND <CONDITION 2>) OR <CONDITION 3> Group together a condition

Mousebindings

General

Scroll Table: Scroll
Chart: Zooms in or out by scrolling up or down respectively
Click Selects the clicked widget, table entry, dialog option, or tab.
Can be disabled via options/flags.

CPU bindings

Scroll Scrolling over an CPU core/average shows only that entry on the chart

Process bindings

Click If in tree mode and you click on a selected entry, it toggles whether the branch is expanded or not

Features

As yet another process/system visualization and management application, bottom supports the typical features:

  • CPU usage visualization, on an average and per-core basis

  • RAM and swap usage visualization

  • Network visualization for receiving and transmitting, on a log-graph scale

  • Display information about disk capacity and I/O per second

  • Display temperatures from sensors

  • Display information regarding processes, like CPU, memory, I/O usage, and process state

  • Process management (well, if process killing is all you need)

It also aims to be:

  • Lightweight

  • Cross-platform - supports 64-bit Linux, Windows, and macOS

In addition, bottom also currently has the following features:

Processes

Process searching

On any process widget, hit / to bring up a search bar. If the layout has multiple process widgets, note this search is independent of other widgets.

search bar image

By default, just typing in something will search by process name:

a simple search

This simple search can be refined by matching by case, matching the entire word, or by using regex:

a slightly better search

Now let's say you want to search for two things - luckily, we have the AND and OR logical operators:

logical operator demo with just ors

logical operator demo with ands and ors

Furthermore, one is able to refine their searches by CPU usage, memory usage, PID, and more. For example:

using cpu filter

You can see all available keywords and query options here.

Process sorting

You can sort the processes list by any column you want by pressing s while on a process widget:

sorting

Tree mode

Use t or F5 to toggle tree mode in a process widget. This is somewhat similar to htop's tree mode.

Standard tree

Sorting works as well, but it is done per groups of siblings. For example, by CPU%:

Standard tree

You can also still filter processes. Branches that entirely do not match the query are pruned out, but if a branch contains an element that does match the query, any non-matching elements will instead just be greyed out, so the tree structure is still maintained:

Standard tree

Zoom

Using the +/- keys or the scroll wheel will move the current time intervals of the currently selected widget, and = to reset the zoom levels to the default. Widgets can hold different time intervals independently. These time intervals can be adjusted using the -t/--default_time_value and -d/--time_delta options, or their corresponding config options.

Expand

Only care about one specific widget? You can go to that widget and hit e to make that widget expand and take up the entire drawing area. You can minimize this expanded widget with Esc or pressing e again.

Basic mode

Using the -b or --basic_mode (or their corresponding config options) will open bottom in basic mode. There are no charts or expanded mode when using this, and tables are condensed such that only one table is displayed at a time.

basic mode image

Note custom layouts are currently not available when this is used.

Config files

bottom supports reading from a config file to customize its behaviour and look. By default, bottom will look at (based on dirs):

OS Location
~/.config/bottom/bottom.toml or $XDG_CONFIG_HOME/bottom/bottom.toml Linux
$HOME/Library/Application Support/bottom/bottom.toml macOS
C:\Users\<USER>\AppData\Roaming\bottom\bottom.toml Windows

Note that if a config file does not exist at either the default location or the passed in location via -C or --config, one is automatically created with no settings applied.

Config flags

The following options can be set under [flags] to achieve the same effect as passing in a flag on runtime. Note that if a flag is given, it will override the config file.

These are the following supported flag config values, which correspond to the flag of the same name described in Flags:

Field Type
hide_avg_cpu Boolean
dot_marker Boolean
left_legend Boolean
current_usage Boolean
group_processes Boolean
case_sensitive Boolean
whole_word Boolean
regex Boolean
show_disabled_data Boolean
basic Boolean
hide_table_count Boolean
use_old_network_legend Boolean
battery Boolean
rate Unsigned Int (represents milliseconds)
default_time_value Unsigned Int (represents milliseconds)
time_delta Unsigned Int (represents milliseconds)
temperature_type String (one of ["k", "f", "c", "kelvin", "fahrenheit", "celsius"])
default_widget_type String (one of ["cpu", "proc", "net", "temp", "mem", "disk"], same as layout options)
default_widget_count Unsigned Int (represents which default_widget_type)
disable_click Boolean
color String (one of ["default", "default-light", "gruvbox", "gruvbox-light"])
mem_as_value Boolean
tree Boolean

Theming

The config file can be used to set custom colours for parts of the application under the [colors] object. The following labels are customizable with strings that are hex colours, RGB colours, or specific named colours.

Supported named colours are one of the following strings: Reset, Black, Red, Green, Yellow, Blue, Magenta, Cyan, Gray, DarkGray, LightRed, LightGreen, LightYellow, LightBlue, LightMagenta, LightCyan, White.

Labels Details Example
Table header colours Colour of table headers table_header_color="255, 255, 255"
CPU colour per core Colour of each core. Read in order. cpu_core_colors=["#ffffff", "white", "255, 255, 255"]
Average CPU colour The average CPU color avg_cpu_color="White"
All CPUs colour The colour for the "All" CPU label all_cpu_color="White"
RAM The colour RAM will use ram_color="#ffffff"
SWAP The colour SWAP will use swap_color="#ffffff"
RX The colour rx will use rx_color="#ffffff"
TX The colour tx will use tx_color="#ffffff"
Widget title colour The colour of the label each widget has widget_title_color="#ffffff"
Border colour The colour of the border of unselected widgets border_color="#ffffff"
Selected border colour The colour of the border of selected widgets highlighted_border_color="#ffffff"
Text colour The colour of most text text_color="#ffffff"
Graph colour The colour of the lines and text of the graph graph_color="#ffffff"
Cursor colour The cursor's colour cursor_color="#ffffff"
Selected text colour The colour of text that is selected scroll_entry_text_color="#ffffff"
Selected text background colour The background colour of text that is selected scroll_entry_bg_color="#ffffff"
High battery level colour The colour used for a high battery level (100% to 50%) high_battery_color="green"
Medium battery level colour The colour used for a medium battery level (50% to 10%) medium_battery_color="yellow"
Low battery level colour The colour used for a low battery level (10% to 0%) low_battery_color="red"

Layout

bottom supports customizable layouts via the config file. Currently, layouts are controlled by using TOML objects and arrays.

For example, given the sample layout:

[[row]]
  [[row.child]]
  type="cpu"
[[row]]
    ratio=2
    [[row.child]]
      ratio=4
      type="mem"
    [[row.child]]
      ratio=3
      [[row.child.child]]
        type="temp"
      [[row.child.child]]
        type="disk"

This would give a layout that has two rows, with a 1:2 ratio. The first row has only the CPU widget. The second row is split into two columns with a 4:3 ratio. The first column contains the memory widget. The second column is split into two rows with a 1:1 ratio. The first is the temperature widget, the second is the disk widget.

This is what the layout would look like when run:

Sample layout

Each [[row]] represents a row in the layout. A row can have any number of child values. Each [[row.child]] represents either a column or a widget. A column can have any number of child values as well. Each [[row.child.child]] represents a widget. A widget is represented by having a type field set to a string.

The following type values are supported:

"cpu" CPU chart and legend
"mem", "memory" Memory chart
"net", "network" Network chart and legend
"proc", "process", "processes" Process table and search
"temp", "temperature" Temperature table
"disk" Disk table
"empty" An empty space
"batt", "battery" Battery statistics

Each component of the layout accepts a ratio value. If this is not set, it defaults to 1.

For an example, look at the default config, which contains the default layout.

Furthermore, you can have duplicate widgets. This means you could do something like:

[[row]]
  ratio=1
  [[row.child]]
  type="cpu"
  [[row.child]]
  type="cpu"
  [[row.child]]
  type="cpu"
[[row]]
  ratio=1
  [[row.child]]
  type="cpu"
  [[row.child]]
  type="empty"
  [[row.child]]
  type="cpu"
[[row]]
  ratio=1
  [[row.child]]
  type="cpu"
  [[row.child]]
  type="cpu"
  [[row.child]]
  type="cpu"

and get the following CPU donut: CPU donut

Disk and temperature filtering

You can hide specific disks and temperature sensors by name in the config file via disk_filter and temp_filter respectively. Regex (regex = true) and case-sensitivity (case_sensitive = true) are supported, but are off by default.

For example, let's say , given this disk list:

Disk filter not ignoring list

I wish to only show disks that follow the form /dev/sda\d+, or /dev/nvme0n1p2:

[disk_filter]
is_list_ignored = false
list = ["/dev/sda\\d+", "/dev/nvme0n1p2"]
regex = true

Disk filter not ignoring list

This would ignore anything that does not match either of these two conditions. If I instead wish to ignore anything that matches this list, then I can set is_list_ignored = true instead:

Disk filter ignoring list

Likewise, I can do something similar for temp_filter:

Temp filter before

If I, say, only wanted to see any entry with the words "cpu" or "wifi" in it, case sensitive:

[temp_filter]
is_list_ignored = false
list = ["cpu", "wifi"]
case_sensitive = true

Temp filter after

Now, flipping to case_sensitive = false would instead show:

Temp filter after with case sensitivity off

Battery

You can get battery statistics (charge, time to fill/discharge, consumption in watts, and battery health) via the battery widget.

Since this is only useful for devices like laptops, it is off by default. You can either enable the widget in the default layout via the --battery flag, or by specifying the widget in a layout:

Battery example

Compatibility

The current compatibility of widgets with operating systems from personal testing:

OS CPU Memory Disks Temperature Processes/Search Networks Battery
Linux
Windows ✓ (seems to have issues with dual batteries)
macOS

FAQ

Please see the FAQ for answers to frequently asked questions.

Contribution

Contribution is always welcome! Please take a look at CONTRIBUTING.md for details on how to help.

Contributors

Thanks to all contributors (emoji key):


Marcin Wojnarowski

💻 📦

Mahmoud Al-Qudsi

💻

Andy

💻

Kim Brose

💻

Sven-Hendrik Haase

📖

Artem Polishchuk

📦 📖

Trung Lê

📦 🚇

dm9pZCAq

📦 📖

Thanks

  • This project is very much inspired by both gotop, its successor ytop, and gtop.

  • Basic mode is heavily inspired by htop's design.

  • This application was written with many, many libraries, and built on the work of many talented people. This application would be impossible without their work. I used to thank them all individually but the list got too large...

  • And of course, thanks to all contributors!