Tags

v0.6.0

v0.6.0 - Date Filtering and UX Improvements

Fixed
- Critical date boundary filtering bug preventing data leakage across reporting periods
- Timeline charts now use appropriate time columns (Start for CPU/GPU, Submit for jobs/users)
- Reports validation now prevents generation for incomplete/future periods
- User-friendly error messages for invalid date ranges and report parameters

Improved
- Daily period charts now show day of week (e.g., "Mon Jan 1")
- Clarified "Number of Submitted Jobs" labels to distinguish from started jobs
- Added visual validation warnings for invalid date ranges
- Removed top 10 limits when using color grouping for better data visibility
- Agent retry logic with increased attempts and timeout for reliability

Documentation
- Added comprehensive dashboard screenshot
- Simplified README structure

v0.5.1

Fix node name display for legacy data with compressed SLURM notation

v0.5.0

Release v0.5.0

Deploy Key System & Security Enhancements

Features:
- Deploy key system with IP tracking and 7-day expiration
- One-command agent setup with automatic key exchange
- Secure cluster details page for key management
- Custom SLURM icon with #09a4d6 background
- 14-day default data collection period
- Admin user management stored in database

Security:
- Moved sensitive API keys to dedicated details page
- Deploy key expiration and single-use enforcement
- IP address tracking for deploy key usage
- Improved key rotation workflow

Agent:
- New 'slurm-dashboard setup' command for automated deployment
- GitLab repository installation support
- Extended data collection window to 14 days
- Streamlined configuration process

v0.4.2

v0.4.2 - Database-backed Admin User Management

- Fixed ".env file not found" error when adding admin users through UI
- Admin emails now stored in database (clusters.json) instead of .env file

- Database-backed storage for both admin and superadmin email lists
- Automatic fallback to environment variables for backward compatibility
- Admin users can now be managed through UI without file system access
- Changes take effect immediately without backend restart

v0.4.1

v0.4.1 - Docker Dev Environment Fixes

- Fixed Docker development environment with default admin credentials (admin/admin)
- Fixed demo cluster generation by including scripts directory in Docker image
- Fixed cluster filtering to properly hide deactivated clusters from dashboard dropdown
- Fixed cluster database import path issue causing all clusters to show regardless of active status

- Added automatic fallback support for both data/ and weekly-data/ directory structures
- Improved cluster filtering to read database directly without cross-package imports
- Added ADMIN_USERS environment variable to docker-compose.yml for local development

v0.4.0

- **Admin Panel**: Full cluster management interface at `/admin/clusters`
  - Add/remove clusters with API key authentication
  - Activate/deactivate clusters
  - Auto-generate configurations from existing data
  - User management
- **Demo Cluster Generation**: Create synthetic test data with realistic patterns
  - 2 years of data, 110k jobs
  - Seasonal patterns and simulated outages
  - One-click generation from admin panel
- **Multi-Cluster Support**: Filter dashboard by active clusters only
- **Auto-Configuration**: Automatically detect nodes, accounts, and partitions

- Fixed DAIC stats (82 nodes, 21 partitions)
- Proper SAML logout handling
- Auto-reload dashboard when clusters change
- Removed "With Hardware" confusing stat
- Clarified "API Submissions" vs total jobs

- Updated README with Docker quickstart
- Removed outdated deployment docs
- Added admin panel documentation
- Cleaned up 486 lines of obsolete docs

- Fixed partition splitting in dropdown
- Fixed node expansion for SLURM notation
- Fixed cluster dropdown not updating on activate/deactivate

v0.1.0