Composure

A terminal tool to audit, optimize, and visualize Docker-Compose stacks in real-time.

Quick Install

pip install composure

Or choose your preferred method below.

What is Composure?

Composure is a TUI (Terminal User Interface) dashboard that helps you:

• Monitor all your Docker containers in real-time
• Detect waste by comparing actual resource usage vs allocated limits
• Visualize networks to see how containers connect to each other
• Control containers directly from the terminal (start, stop, restart, view logs)
• Pull images with a single overall progress bar across all images

Features

• Resource Monitoring: See CPU and memory usage for all containers
• Waste Detection: Identify over-provisioned containers with waste scores
• Limit Awareness: Quickly spot containers without resource limits
• Network Visualization: Tree view showing container network topology
• Container Controls: Start, stop, restart containers with keyboard shortcuts
• Live Logs: View recent logs for any container
• Multi-Container Logs: View merged logs from ALL containers with color-coded prefixes (press L)
• Image Pull Progress: Pull all images from docker-compose.yml with a single overall progress percentage
• Parallel Loading: Fast startup even with many containers

Installation

pip (recommended)

pip install composure

pipx (isolated environment)

pipx install composure

uv

uv tool install composure

Docker

docker run -it -v /var/run/docker.sock:/var/run/docker.sock jamesdimonaco/composure

Debian/Ubuntu

# One-line install script
curl -fsSL https://jamesdimonaco.github.io/composure/install.sh | sudo bash

# Add the GPG key
curl -fsSL https://jamesdimonaco.github.io/composure/gpg.key | sudo gpg --dearmor -o /usr/share/keyrings/composure.gpg

# Add the repository
echo "deb [signed-by=/usr/share/keyrings/composure.gpg] https://jamesdimonaco.github.io/composure stable main" | sudo tee /etc/apt/sources.list.d/composure.list

# Install
sudo apt update
sudo apt install composure

From source

git clone https://github.com/JamesDimonaco/composure.git
cd composure
pip install -e .

Usage

composure

Pull Images

Pull all images from your docker-compose.yml with a single overall progress percentage:

composure pull

This shows real-time progress across all images downloading in parallel:

Pulling 3 images from docker-compose.yml

⠙ Overall              ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━  36%  847MB / 2.3GB
    nginx:alpine       ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100%  done
    redis:alpine       ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━  45%  pulling
    postgres:15-alpine ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━  20%  pulling

You can also press p in the TUI to pull images.

JSON Output

For scripting and CI/CD pipelines, use --json to get machine-readable output:

composure pull --json

Outputs one JSON object per line with progress updates:

{"percent": 36.5, "downloaded_bytes": 847000, "total_bytes": 2300000, "images_complete": 1, "images_total": 3, ...}

Python API

Use composure programmatically in your Python scripts:

from composure import pull

# With progress updates
for progress in pull("/path/to/project"):
    print(f"{progress.percent:.0f}% complete")
    print(f"Images: {progress.images_complete}/{progress.images_total}")

# Or just get final result
result = pull("/path/to/project", progress=False)
print(f"Pulled {result.images_complete} images")

The PullProgress object contains:

• percent - Overall percentage (0-100)
• downloaded_bytes / total_bytes - Byte counts
• images_complete / images_total - Image counts
• completed_layers / total_layers - Layer counts
• images - Dict with per-image details

Keyboard Shortcuts

Understanding the Dashboard

Main View

Container     Status   CPU %  CPU Limit  RAM Used  RAM Limit  Efficiency  Waste
nginx         running  0.5%   2 cores    45 MB     512 MB     LOW         90
api           running  2.1%   No limit   128 MB    256 MB     MEDIUM      40
postgres      running  1.2%   1 core     256 MB    512 MB     GOOD        20

Detail Panel

Select a container to see detailed info at the bottom:

nginx  running  Image: nginx:alpine
ID: a1b2c3d4  Ports: 8080:80/tcp, 443:443/tcp  Restart: always
Mounts: ./html:/usr/share/nginx/html
Networks: frontend, backend
Logs: (press 'l' for full, 'L' for all)
  2024-01-15T10:30:45 GET /api/health 200

Columns Explained

Waste Score

The waste score helps identify containers that have been allocated far more resources than they're using:

Network View

Press n to see a tree view of your Docker networks:

Docker Networks
├── Compose Networks
│   └── myapp_default (3 containers)
│       ├── nginx
│       ├── api
│       └── postgres
└── System Networks
    └── bridge (empty)

Requirements

• Python 3.9+
• Docker Engine running locally
• Access to Docker socket

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

MIT License - see LICENSE for details.

Links

• PyPI Package
• Docker Hub
• GitHub Repository
• Report Issues