Patterns of SaaS: Refactoring Databases
summary
Application runs across 6000 schemas with 76M transactions per week and performs half a million schema changes overnight by sequencing small, re-runnable DDL/DML files (numbered by tens) and running them via a homegrown shell automation tool that writes into a run_script table. Releases use pre/main/post scripts, automated reviews and backups of affected data, QA integration tests, and multiple-engineer SQL reviews. Result: near-zero downtime rollouts, point-of-no-return commits, and a reduction of release windows from hours to about 15 minutes.
AI-written summary · about
chapters (26)
- 0:00Start
- 0:40Why Are We Here Today Database Refactoring
- 2:11Introductions
- 5:01Database Refactoring
- 14:08Key Steps
- 15:58Structural Type
- 16:13Data Quality Refactorings
- 16:37Referential Integrity
- 17:08Functional
- 17:35The Non Refactoring Transformations
- 17:46What Are the Most Common Structural Refactoring Types
- 19:46The Sinking Trigger
- 20:18Data Migration
- 21:29Data Quality Refactoring Types
- 23:20Referential Integrity Validation
- 23:42Deferred Constraint Validation
- 24:41Testing
- 25:15Migration Testing
- 28:01Manage all Database Artifacts and Source Control
- 32:41Run Script Table
- 33:47Common Dml Patterns
- 38:12Sql Review
- 44:02Schema Inconsistencies
- 44:21Data Inconsistency
- 47:19Key Messages
- 50:33Audit Process
related talks
33mDatabase Schema Migrations with Zero Downtime
31mHelp! My Database Is The Bottleneck
31mDevOps Practices for the Database Team
37mDatabase Schema Migrations with Zero Downtime
21mApplying Continuous Delivery to Database Schema
43mDatabase in a SaaS application
46mDatabase DevOps - It is Possible!
45mEnabling Continuous Delivery with Database Practices