🦆 DUKS - Dashboard for Unified Kernel Statistics

How is the Linux Kernel development going ? Duks is a [work in progress] data visualization tool to help with that.

duks.rcpassos.me

Development

To run DUKS in development, it is recommended to use the dev-compose.yaml file. This starts the server with the Backend (that also serves the Frontend).

podman-compose -f dev-compose.yaml up

Obtain the sample data files in the archived replication package.

The file structure

├── data    -> where the data files should be stored
├── nginx    -> configurations for nginx (for production)
├── scripts   -> scripts used to generate the dataset
└── server   -> backend source code
    └── static -> static Frontend files

Running from zero

Bootstrapping this from zero requires a few external resources.

Requirements:

• Local compressed Software Heritage Graph (at least “History and hosting” graph: 1.4TiB for 2024-12-06) : https://docs.softwareheritage.org/devel/swh-export/graph/index.html
• Linux kernel mainline tree
• Software heritage GRPC server: https://docs.softwareheritage.org/devel/swh-graph/grpc-api.html#starting-the-server

Preparing the dataset

This repository is organized to use uv as its python dependencies, this document will use it. Run scripts from the project root directory (as it will create a ./data/ folder relative to the scripts). Other options should work.

1. run the GRPC server with the local graph, following the upstream guide. Example: swh-graph-grpc-serve --bind 0.0.0.0:50091 /media/research/2025-05-18-history-hosting/graph
2. run uv run scripts/grpc_script.py to execute the BFS on the graph. This script will look for the last commit in the DEFAULT_BRANCH for the chosen KERNEL_TREE. Defaults to the master in the mainline kernel. It is also possible to pre-load commits manually. Must set the GRAPH_GRPC_SERVER and INITIAL_NODE, if non-default port or commit would be used.
3. run the uv run scripts/enrich_from_git.py script, pointing to the mainline branch path. This will load data that is unavailable in the graph. Tags can be loaded from the previous step, or loaded here if LOAD_TAGS_FROM_REPO is set. Example: KERNEL_PATH=/media/research/linux uv run scripts/enrich_from_git.py
4. run the scripts/get_official_kernel_maintainers.py to read the contents of the maintainers file in all its changes. This step is independent from others, besides step 5. Example: KERNEL_PATH=/media/research/linux uv run scripts/get_official_kernel_maintainers.py
5. run the [scripts/stitch_data_into_final_payload.py] to get the daily output (with some calculation) and with all files in a single v

Running application

Run either podman-compose -f dev-compose.yaml up for development or podman-compose up for production build

Research

Papers related to this software:

• Streamlining Analyses on the Linux Kernel with DUKS
• DUKS: visualizações e análises unificadas para o Kernel Linux 🇧🇷

PASSOS, Rafael; PILONE, Arthur; TADOKORO, David; MEIRELLES, Paulo