Set Admin Performance Overrides
Set Admin Performance Overrides
Overview
Flow ID: set-admin-overrides
Category: Settings & Configuration
Estimated Duration: 1-2 minutes
User Role: Power Users / Administrators
Complexity: Simple
Purpose: Configure advanced system-wide performance settings including CPU worker count for parallel processing and default chat template. These settings affect all users and all operations.
Trigger
What initiates this flow:
- User manually initiates
Specific trigger: Administrator wants to optimize system performance or set organizational defaults.
Prerequisites
Before starting, users must have:
- Understanding of performance implications
- Knowledge of system hardware (CPU cores)
- Admin access or confidence to change system settings
User Intent Analysis
Primary Intent
Optimize system-wide performance by configuring worker counts and default templates.
Secondary Intents
- Maximize hardware utilization
- Standardize workflows with default templates
- Fine-tune for specific use cases
Step-by-Step Flow
Main Path (Happy Path)
Step 1: Navigate to Admin Overrides
- User Action: Settings > Admin Overrides tab
- System Response: Admin settings page displays
- UI Elements Visible:
- “Admin Overrides” tab highlighted
- Warning banner: “Admin settings affect all users and operations”
- Performance section
- Workflow section
Step 2: Review Current Settings
- User Action: Note current values
- UI Elements Visible:
- Performance Section:
- “Additional CPU Workers” input field
- Current value shown (e.g., “8”)
- Available cores display: “Available CPU Cores: 16”
- Workflow Section:
- “Default Chat Template” dropdown
- Current selection shown
- Performance Section:
Step 3: Adjust CPU Workers
- User Action: Change number in CPU workers field
- System Response:
- Value updates
- Validation ensures within range (1 to max cores)
- UI Elements Visible:
- Number input showing new value
- Range hint: “1 to 16” or similar
- Recommendation: Set to number of available cores minus 2-4 for system overhead
Step 4: Select Default Template (Optional)
- User Action: Open default template dropdown and select template
- System Response:
- Template selected
- This template will be used when creating new chats
- UI Elements Visible:
- Dropdown with template options:
- “No Default Template”
- Available templates listed
- Dropdown with template options:
Step 5: Save Changes
- User Action: Click “Save” button (if not auto-save)
- System Response:
- Settings saved
- Success notification
- Changes take effect
- UI Elements Visible:
- Success message
- Saved values confirmed
Final Step: Admin Settings Configured
- Success Indicator:
- Settings saved successfully
- Take effect immediately for new operations
- System State Change:
- Worker count updated globally
- Future embedding operations use new worker count
- New chats use selected default template
- Next Possible Actions:
- Test performance with new settings
- Run jobs to verify worker count optimization
- Create chats to verify template behavior
Error States & Recovery
Error 1: Worker Count Too High
Cause: Set higher than available cores
User Experience:
- Warning or auto-constrains to maximum
Recovery Steps:
- System prevents invalid values
- Set to reasonable number (cores minus 2-4)
Pain Points & Friction
Identified Issues:
No Guidance on Optimal Worker Count
- Impact: Users don’t know best value
- Potential Improvement: Auto-optimize button, recommendations
Changes Affect All Operations
- Impact: No per-user or per-job customization
- Potential Improvement: Per-user profiles or job-specific overrides
Design Considerations
Following Contextual Design Principles:
- Automation Opportunities: Auto-detect optimal worker count
- User Trust: Clear warning about system-wide impact
- Cognitive Load: Simple number input
Related Flows
- Run Full Benchmark Suite - Test worker configuration
- Create New Blockify Job - Uses worker settings
Technical References
Knowledge Base Sections:
- src/components/admin-overrides-tab/index.js - Admin settings UI
- src/engines/bulk-engines/bulk-embedding.js - Worker management
Version History
| Date | Version | Author | Changes |
|---|---|---|---|
| 2025-10-04 | 1.1 | Iternal Technologies | Initial documentation |
Notes
Worker Count Recommendations:
- Conservative: cores - 4 (leaves room for system)
- Balanced: cores - 2
- Aggressive: cores - 1 (may impact responsiveness)
Best Practices:
- Test with benchmarks after changing
- Monitor system responsiveness
- Leave headroom for OS and other apps
Common User Questions:
- “How many workers should I use?” - Start with cores minus 2
- “Will more workers always be faster?” - Not always; diminishing returns and overhead
- “Can I change per job?” - No, system-wide setting