formbricks-formbricks/.cursor/rules/database-performance.mdc

---
description: 
globs: 
alwaysApply: false
---
# Database Performance & Prisma Best Practices

## Critical Performance Rules

### Response Count Queries
- **NEVER** use `skip`/`offset` with `prisma.response.count()` - this causes expensive subqueries with OFFSET
- Always use only `where` clauses for count operations: `prisma.response.count({ where: { ... } })`
- For pagination, separate count queries from data queries
- Reference: [apps/web/lib/response/service.ts](mdc:apps/web/lib/response/service.ts) line 654-686

### Prisma Query Optimization
- Use proper indexes defined in [packages/database/schema.prisma](mdc:packages/database/schema.prisma)
- Leverage existing indexes: `@@index([surveyId, createdAt])`, `@@index([createdAt])`
- Use cursor-based pagination for large datasets instead of offset-based
- Cache frequently accessed data using React Cache and custom cache tags

### Date Range Filtering
- When filtering by `createdAt`, always use indexed queries
- Combine with `surveyId` for optimal performance: `{ surveyId, createdAt: { gte: start, lt: end } }`
- Avoid complex WHERE clauses that can't utilize indexes

### Count vs Data Separation
- Always separate count queries from data fetching queries
- Use `Promise.all()` to run count and data queries in parallel
- Example pattern from [apps/web/modules/api/v2/management/responses/lib/response.ts](mdc:apps/web/modules/api/v2/management/responses/lib/response.ts):
```typescript
const [responses, totalCount] = await Promise.all([
  prisma.response.findMany(query),
  prisma.response.count({ where: whereClause }),
]);
```

### Monitoring & Debugging
- Monitor AWS RDS Performance Insights for problematic queries
- Look for queries with OFFSET in count operations - these indicate performance issues
- Use proper error handling with `DatabaseError` for Prisma exceptions