clusterscope

What it is and what it does

Clusterscope is a tool for HPC administrators and researchers who need to programmatically or interactively query cluster state. It provides both a Python library and a command-line interface to extract information about nodes, GPUs, CPUs, memory, and job configurations from Slurm-managed clusters. The library can be imported for scripted cluster introspection, while the CLI (`cscope`) offers immediate access to common queries like node resource counts, GPU availability, and AWS environment checks.

The package is lightweight, depending only on click for CLI scaffolding, and targets modern Python (>=3.10). It is actively maintained by Facebook Research and suitable for automation, monitoring integration, and cluster capacity planning workflows where you need to query live cluster state without writing raw Slurm commands.

Use it for:

• Query available GPU types and counts across cluster nodes before submitting distributed training jobs
• Programmatically check CPU and memory per node to generate job submission requirements dynamically
• Verify AWS environment and NCCL settings when running on AWS HPC clusters
• Integrate cluster metadata into monitoring dashboards or resource allocation scripts
• Automate cluster health checks and node inventory reporting in CI/CD or deployment pipelines

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Clusterscope extracts and queries information from HPC clusters and jobs via a CLI tool and Python library, with built-in support for Slurm cluster metadata and AWS/NCCL diagnostics.

Yes, if you manage or work with Slurm-based HPC clusters and need programmatic access to cluster metadata. The low install friction, active maintenance, permissive license, and clean API make it a practical choice for cluster automation and job submission workflows. Not applicable if your cluster uses a different scheduler or if you only need occasional manual cluster queries.

Install

clusterscope on PyPI

pip install clusterscope

uv add clusterscope

poetry add clusterscope

Installing clusterscope

Before you install

Low install friction with only two lightweight runtime dependencies (click and click-option-group). The package is actively maintained with recent commits and a stable release cadence since mid-2025.

License in practice

MIT license permits unrestricted use, modification, and distribution with minimal restrictions—suitable for both open-source and proprietary projects.

Quickstart

pip install clusterscope

import clusterscope
print(clusterscope.cluster())

# Or use the CLI:
# cscope info
# cscope gpus
# cscope cpus

Requires Python >=3.10. Designed for HPC environments with Slurm; functionality depends on cluster availability and proper Slurm configuration.

Verify before relying

• Whether the package works on non-Slurm HPC schedulers (e.g., PBS, LSF) or is Slurm-specific
• Performance characteristics when querying large clusters with many nodes or jobs
• Exact AWS/NCCL diagnostic capabilities and whether they require AWS credentials or specific instance types

Package facts

Evidence: clusterscope-0.0.32-py3-none-any.whl

Tags

More Distributed Computing packages