From bottlenecks to breakthroughs: Dutchie’s database migration journey

This article details how Dutchie, a cannabis industry technology platform, successfully migrated its database infrastructure from Azure SQL Managed Instance to Amazon RDS for SQL Server to handle peak traffic demands.

• Dutchie processes $22 billion in annual transaction volume across 6,500+ cannabis businesses
• Migration driven by need to scale dynamically, optimize performance, reduce costs, and prepare for 4/20 week peak traffic
• Previous Azure SQL Managed Instance had rigid SKU configurations limiting IOPS and scaling flexibility
• Implemented x2iedn RDS instances with high core-to-memory ratios and local NVMe storage for tempdb
• Enabled Read Committed Snapshot Isolation (RCSI) to reduce blocking under high-concurrency workloads
• Automated migration using Python scripts and AWS CLI for zero-data loss cutover within two-hour window
• Achieved improved I/O latency, cache efficiency, and independent scaling of compute and storage
• System successfully handled highest sustained traffic week in company history during 4/20 2025
• Key success factors: baseline performance measurement, automation, rehearsals with production-sized datasets, and close AWS collaboration

Dutchie's migration demonstrates how RDS for SQL Server provides the flexibility and performance needed to support mission-critical, high-volume transaction workloads during peak demand periods.