jupyterlab-resource-monitor-extension 1.0.27

JupyterLab extension replacing jupyter-resource-usage with a less intrusive, better implementation of resource usage monitoring

jupyterlab_resource_monitor_extension

JupyterLab extension that replaces jupyter-resource-usage with a less intrusive and better implementation of resource usage monitoring.

This extension is composed of a Python package named jupyterlab_resource_monitor_extension for the server extension and a NPM package named jupyterlab_resource_monitor_extension for the frontend extension.

Features

• Memory monitor in the status bar - current memory usage with the maximum available memory shown alongside; under cgroup v2 this is the server's own cgroup, and without it the reading is host-wide
• CPU usage in the status bar - live CPU utilization next to the memory indicator, as a percentage, a compact bar, or a grid; same scope as the memory reading
• Per-core grid - the grid mode draws the cores assigned to this server as a square of cells, each gaining opacity with its own load and burning brighter past 75%, so a few pinned cores and an evenly busy machine look different at the same percentage
• Kernel usage panel - sidebar panel showing memory and CPU for the kernel attached to the notebook you are looking at
• Container aware - reads the cgroup v2 limits when present, so the ceiling shown is the container's limit rather than the host's RAM
• Event-loop safe metrics collection - metrics are gathered off the tornado event loop, so polling never blocks websockets, kernel messages, or static file serving
• Bounded collection cost - no synchronous recursive walk of the server's child process tree, so a large or leaked process population cannot degrade JupyterLab responsiveness
• Measures kernels only while you are watching - the kernel panel samples solely while it is open, and stops entirely when it is closed
• Drop-in replacement - designed to replace jupyter-resource-usage with a safer implementation

How the collection stays off the event loop, what bounds its cost, and how it degrades is described in How Resource Collection Avoids Blocking the Server.

Settings

Configure under Settings -> Settings Editor -> Resource Monitor.

Setting	Default	Effect
showMemory	true	Show the memory item in the status bar
showCpu	true	Show the CPU item in the status bar
showMemoryPercent	false	Append the used-of-limit percentage to the memory item
cpuMode	normalised	normalised shares all cores (0-100%), aggregate sums each core's own 100%, bar draws a filled bar, grid draws the cores as a square
showKernelUsage	true	Show the kernel usage sidebar panel

Hovering either status bar item shows the full detail regardless of these settings, including the memory percentage and the CPU core count.

Requirements

• JupyterLab >= 4.0.0
• Linux host - the sampler reads cgroup v2 and /proc; on macOS and Windows no reading is available, so the status bar items stay hidden and the kernel panel reports Measuring...

Install

To install the extension, execute:

pip install jupyterlab_resource_monitor_extension

Uninstall

To remove the extension, execute:

pip uninstall jupyterlab_resource_monitor_extension