nuse 0.1.2

No project description provided

nuse

nuse is a simple resource monitoring tool for SLURM jobs. It allows users to view the CPU and memory usage logs for individual jobs and the entire node. With nuse, when running a SLURM-based workflow, you can easily monitor resource consumption through command-line tools.

Features

• Job-Specific Monitoring: Capture and view resource usage for each job (e.g., using cgroup filtering).
• Node-Wide Monitoring: Automatically collect a separate log for overall node usage.
• Command-Line Interface: Installed via pip, the nuse command lets you quickly view logs with a simple one-liner.
• Custom Log Directory: Easily configure where logs are stored by setting the MONITOR_LOG_DIR environment variable.

Installation

Install nuse directly from PyPI:

pip install nuse

Usage

In your script include

from nuse import start_monitoring

start_monitoring(filter_cgroup=True)

# your code

Job-Specific Log: To display the resource usage log for a specific job on a node, run in your CLI:

nuse node305 49847516

Here, node305 is the node's short name and 49847516 is the SLURM job ID.

Configuration

Log Directory:

• By default, nuse stores logs in the directory ~/.monitoring. To change the log directory, set the environment variable before running your jobs:

export MONITOR_LOG_DIR="/path/to/your/log_directory"

How it Works

When included in your SLURM job pipeline (via start_monitoring(filter_cgroup=True) from the nuse package), nuse will:

• Create a job-specific log file with a naming convention like cpu_memory_usage_.cluster_<SLURM_JOB_ID>.log.
• Also create a node-wide log file named cpu_memory_usage__node.log for monitoring overall resources.

Then, the nuse CLI tool uses these logs to provide an easy-to-read, real-time view of resource usage.

Happy monitoring!