Production Data Sync

The Sync medium copies data from one environment to another with built-in PII anonymization. Debug with realistic data without exposing personal information.

How It Works

# Sync with anonymization
confiture sync --from production --to local --anonymize

✓ Syncing production → local
  → Anonymizing PII columns...
  → tb_user: 12,450 rows (6,500 rows/sec with anonymization)
  → tb_post: 45,230 rows (70,000 rows/sec)
  → tb_comment: 128,900 rows (70,000 rows/sec)
✓ Sync complete in 4.2s

Commands

# Full sync with anonymization
confiture sync --from production --to local --anonymize

# Sync specific tables only
confiture sync --from production --to staging --tables users,posts

# Sync without anonymization (staging to staging)
confiture sync --from staging --to local

# Resume an interrupted sync
confiture sync --resume --checkpoint sync.json

# Dry run — show what would be synced
confiture sync --from production --to local --anonymize --dry-run

Anonymization Strategies

Configure anonymization rules in confiture.yaml:

confiture.yaml

sync:
  anonymize:
    tb_user:
      email: email        # alice@example.com → user_a1b2c3@example.com
      name: name          # Alice Johnson → User A1B2
      phone: phone        # +1-555-1234 → +1-555-4567
      bio: redact         # Any value → [REDACTED]
      password_hash: hash # Consistent hash for testing
    tb_payment:
      card_number: redact
      billing_address: redact

Available Strategies

Strategy	Input	Output	Use Case
email	alice@example.com	user_a1b2c3@example.com	Email fields
name	Alice Johnson	User A1B2	Name fields
phone	+1-555-1234	+1-555-4567	Phone numbers
redact	Any value	[REDACTED]	Sensitive text, addresses
hash	Any value	Consistent hash	Passwords, tokens (preserves uniqueness)

Anonymization Properties

• Deterministic — the same input always produces the same output (useful for foreign key consistency)
• Non-reversible — you cannot recover the original data from the anonymized version
• Format-preserving — emails look like emails, phones look like phones

GDPR and CCPA Compliance

The non-reversible, format-preserving anonymization makes Sync suitable for privacy-compliant data operations under GDPR and CCPA.

Performance

Mode	Throughput	Use Case
Without anonymization	~70,000 rows/sec	Staging-to-staging, non-sensitive data
With anonymization	~6,500 rows/sec	Production-to-local, PII data

For large syncs (millions of rows), use checkpoints to resume if interrupted:

# Start sync with checkpoint file
confiture sync --from production --to local --anonymize --checkpoint sync.json

# Resume if interrupted
confiture sync --resume --checkpoint sync.json

When to Use Sync

Use Sync for

• Local debugging with realistic data volumes and patterns
• Staging performance testing with production-scale data
• Reproducing production bugs locally
• Privacy-compliant data operations (GDPR, CCPA)

Don't use Sync for

• Creating fresh databases — use Build instead
• Schema changes — use Migrate instead

Next Steps

Confiture Overview

All 4 Mediums and when to use each one.

Build from DDL

For fresh databases — fast setup from DDL files in under 1 second.

Incremental Migrations

For schema evolution on existing databases with data.

Schema-to-Schema

Zero-downtime refactoring for major production changes.