sf-agent-framework/sf-core/tasks/ldv-optimization.md

AI Agent Orchestration Framework for Salesforce Development - Two-phase architecture with 70% context reduction

# LDV Optimization

## Purpose

Optimize Salesforce implementations for Large Data Volumes (LDV) to ensure
performance, scalability, and platform stability.

## Instructions

1. **Data Volume Assessment**
   - Analyze current data volumes
   - Project growth rates
   - Identify large objects
   - Review data distribution
   - Assess query patterns

2. **Performance Analysis**
   - Profile slow queries
   - Identify non-selective filters
   - Review sharing calculations
   - Analyze report performance
   - Check integration impacts

3. **Indexing Strategy**
   - Design custom indexes
   - Optimize standard indexes
   - Plan composite indexes
   - Consider skinny tables
   - Review index usage

4. **Data Architecture Optimization**
   - Implement data archiving
   - Design efficient relationships
   - Optimize field types
   - Reduce formula fields
   - Minimize roll-up summaries

5. **Query Optimization**
   - Improve SOQL selectivity
   - Optimize filter conditions
   - Reduce query complexity
   - Implement query pagination
   - Use aggregate queries efficiently

6. **Platform Features**
   - Configure Big Objects
   - Implement Platform Cache
   - Use External Objects
   - Enable Skinny Tables
   - Configure Custom Indexes

## Input Requirements

- Data volume metrics
- Query performance logs
- Object relationships
- Usage patterns
- Growth projections
- Performance requirements

## Output Format

- LDV Optimization Plan:
  - Volume analysis report
  - Performance assessment
  - Optimization recommendations
  - Implementation roadmap
  - Monitoring approach
  - Success metrics

## Optimization Techniques

- **Data Model**: Denormalization, archiving
- **Indexing**: Custom indexes, skinny tables
- **Queries**: Selectivity, pagination
- **Sharing**: OWD optimization, ownership
- **Processing**: Batch, async, queueable
- **Caching**: Platform cache, custom cache

## Performance Thresholds

- Query timeout: 120 seconds
- Index selectivity: <10% of records
- Sharing recalculation: <2M records
- Report timeout: 10 minutes
- API timeout: varies by type

## Best Practices

- Design for scale from start
- Archive historical data
- Monitor continuously
- Test with production volumes
- Use selective queries
- Optimize sharing model
- Leverage platform features