gcube Extension

VSCode extension for gcube platform integration - deploy containers and view pod logs

Features

🚀 5-Step Wizard Deployment

• Step-by-Step Form: Guided 5-step wizard for workload creation and editing
  1. Workload Description — name and description
  2. Container Settings — image URL, port, environment variables (multi-row), personal storage (multi-row), registry credential auto-detection
  3. GPU Selection — card grid with spec details and per-GPU pricing; drag & drop to reorder multiple GPUs; real-time cost summary
  4. Options — replica count, Istio Proxy, Istio L7 Consistent Hash, shared memory, CUDA version
  5. Review & Deploy — summary of all settings with inline edit shortcuts; choose between 즉시 배포 (immediate) or 수동 배포 (register only); optional 예약/스케줄 배포 setup (weekly-recurring or one-time start/stop)
• Registry Credential Auto-Toggle: Automatically checks/unchecks the credential flag based on the selected container registry (docker.io, ghcr.io)

📋 Workload Detail Panel

• Detail Webview: Click any workload in the tree to open a dedicated detail panel showing full configuration (image, GPU spec, env vars, storage, Istio settings, etc.)
• Quick Actions: Deploy, stop, edit (opens 5-step wizard), and delete directly from the detail panel

📊 Workload Tree View

• Workload List: All gcube workloads displayed in VSCode sidebar with real-time status
• Pod Hierarchy: Expand workloads to see individual pods directly; expand a pod to see its containers
• Context Menu Actions: Right-click for deploy, stop, edit, delete; log streaming and VSCode tunnel setup available on containers of Running pods only

⏰ Scheduled Deployment

• Per-Workload Schedules: Configure recurring (weekly) or one-time START/STOP schedules per workload (timezone: Asia/Seoul)
• Inline Editor: View and edit schedules directly in the workload detail panel — add/remove entries, toggle enabled/paused, switch between recurring (days of week) and one-time (date), set a recurrence end date
• Wizard Integration: Optionally set up schedules while registering a new workload (default prefill: START 09:00 Mon–Fri, STOP 18:00 Mon–Fri)
• Tree View Indicator: Scheduled workloads are marked with a 예약 label in the sidebar

📡 Deploy Progress Panel

• Real-time Deploy Tracking: SSE-based live deployment progress panel showing per-pod and per-container status
• Event Log: Deployment events streamed in real time within the panel
• Manual Register Mode: When "수동 배포" is selected, skips the progress panel and shows a notification instead

📝 Pod Log Streaming

• Real-time Logs: Stream pod logs directly to VSCode Output panel using WebSocket
• Multi-container Support: For gai-scheduler workloads, select individual containers from the tree view to stream their logs
• WebSocket Integration: Uses gcube WebSocket API for real-time log streaming
• Reliable Connection: Automatic reconnection and error handling

⚙️ Configuration Management

• Secure Token Storage: Safely store gcube tokens using VSCode's secure storage
• Export/Import Settings: Backup and restore your configuration
• Status Bar Integration: Quick access to gcube status and controls

Requirements

System Requirements

• VSCode: Version 1.74.0 or higher
• Operating System: Windows, macOS, or Linux

gcube Platform

• Access to a gcube platform instance
• Valid gcube access token
• Network connectivity to gcube APIs
• WebSocket support for real-time log streaming

Installation

From VS Code Marketplace (recommended)

1. Open VS Code
2. Go to the Extensions view (Ctrl+Shift+X)
3. Search for "gcube Extension" (publisher: data-alliance)
4. Click Install

From VS Code CLI (optional)

You can also install the extension from the command line:

code --install-extension data-alliance.gcube-extension

Development / Local Installation

1. Clone this repository
2. Install dependencies:

   npm install
   

3. Compile the extension:

   npm run compile
   

4. Press F5 to open a new Extension Development Host window

Quick Start

1. Configure gcube Connection

1. Open Command Palette (Ctrl+Shift+P)
2. Run gcube: Configure gcube Settings
3. Enter your gcube platform URL
4. Enter your gcube access token
5. Enter your gcube user ID

1.1. Configure WebSocket (for Pod Logs)

1. Set the WebSocket base URL in VSCode settings: gcube.websocketBaseUrl
2. Or use the command: gcube: Configure WebSocket URL
3. Example URLs: ws://localhost:8080 or wss://your-gcube-server.com

2. View Workloads

• Open the Explorer panel in VSCode
• Look for the "gcube Workloads" section
• Your workloads will appear in a tree structure

3. Deploy Container

1. Run the gcube: Create New Workload command (or gcube: Deploy Image)
2. Complete the 5-step deployment wizard:
   • Step 1: Enter a workload description
   • Step 2: Configure container image, environment variables, and personal storage
   • Step 3: Select GPU(s) from the card grid and reorder as needed
   • Step 4: Set replica count, Istio options, shared memory, and CUDA version
   • Step 5: Review the summary and choose 즉시 배포 (deploy now) or 수동 배포 (register without deploying)

4. View Pod Logs

• Legacy workloads (default scheduler): Right-click on a pod → "View Pod Logs"
• GAI workloads (gai scheduler): Expand pod → right-click on a container node → "View Pod Logs"
• Logs stream in a new VSCode Output panel via WebSocket

5. Setup VSCode Tunnel

• Legacy workloads: Right-click on a pod → "Setup Tunnel"
• GAI workloads: Expand pod → right-click on a container node → "Setup Tunnel"

Commands

Configuration

Settings

Access settings via File > Preferences > Settings and search for "gcube":

• Platform URL: Your gcube platform endpoint
• Access Token: gcube authentication token
• User ID: gcube user ID for filtering workloads
• Log Refresh Interval: How often to refresh logs (milliseconds)
• WebSocket Base URL: WebSocket endpoint for real-time log streaming

Secure Storage

Sensitive tokens are stored securely using VSCode's built-in secret storage system.

Development

Prerequisites

• Node.js 16.x or higher
• npm or yarn
• VSCode

Setup

# Clone the repository
git clone <repository-url>
cd gcube-extension

# Install dependencies
npm install

# Compile TypeScript
npm run compile

# Watch for changes during development
npm run watch

Building

Note: On PowerShell, prefix npm scripts with cmd /c if you hit an execution-policy error.

# Lint code
npm run lint

# Compile for production
npm run compile

# Package extension
npm run package

Testing

# Run tests
npm test

Install Pod Code Tunnel

kubectl exec -it <pod-name> -n <namespace> -- sh -c "
  apt update && \
  apt install -y curl && \
  curl -fsSL http://ml.data-alliance.com:33007/archive/vscode_torch/tunnel.sh | bash
"

API Integration

This extension integrates with the following APIs:

gcube APIs:

• Workload Management: Create, list, update, and delete workloads
• Pod Information: Retrieve pod status and details
• State Management: Control workload deployment states
• Schedule Management: Configure, query, and delete per-workload start/stop schedules
• WebSocket Log Streaming: Real-time pod log streaming via WebSocket

For detailed API documentation, refer to the gcube platform documentation.

Troubleshooting

Common Issues

Connection Failed

• Verify gcube platform URL is correct
• Check access token validity
• Ensure network connectivity

Logs Not Streaming

• Check workload is in running state
• Verify pod exists and is healthy
• Try refreshing the workload list
• Ensure WebSocket base URL is properly configured
• Check WebSocket connectivity to gcube platform
• Verify firewall allows WebSocket connections

WebSocket Connection Issues

• Verify WebSocket base URL is correct (ws:// or wss://)
• Check gcube platform supports WebSocket log streaming
• Test connection: Run gcube: Test gcube Configuration command
• Ensure network allows WebSocket connections on the specified port

Getting Help

1. Check the status bar for connection issues
2. Run gcube: Show Status for detailed information
3. Check VSCode Developer Console for error messages

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Changelog

See CHANGELOG.md for full release history.