Feature management: De-risking migrations and releases for PowerSchool
summary
PowerSchool describes an incremental refactor that pulls UI into micro frontends, breaks a monolith into nine domain services, and migrates tenant data with AWS Lambda, S3, EventBridge and Step Functions into DynamoDB, all coordinated with LaunchDarkly feature flags. The team reports deleting about 150,000 lines of legacy code, serving ~10 million Naviance students (45M across PowerSchool), cutting feature rollout time to minutes, and using LaunchDarkly’s migration flags and release pipelines to stage, validate and rollback per-customer migrations.
AI-written summary · about
chapters (6)
- 0:00Introduction and setup
- 5:46Why releases are risky
- 14:02Breaking monolith into services
- 26:29Four-step release strategy
- 39:26Demo of feature management
- 53:20Q&A and closing
related talks
28mEnable processing capacity and scale up with AWS Lambda
49mBuilding and operating at scale with feature management
27mMicro-transformation
07mScaling with LaunchDarkly: A feature toggle story!
29mMoving from A and B to ~150 microservices, the journey and learnings
11mOptimizing Product Management with Multi-Tenancy Architecture
01h 01mLegacy Architecture Migration Patterns with DDD
30mThe transformative power of modern applications