Overview
Watch a 23-minute conference talk from AWS re:Invent 2023 exploring Netflix's development of Maestro, their innovative workflow orchestrator designed for managing large-scale data and machine learning operations in the cloud. Discover how Netflix addresses the growing demands of big data and ML by creating a system that enhances scalability, reliability, and usability. Learn about Maestro's capabilities in handling complex workflows with hundreds of thousands of nested jobs while maintaining comprehensive lineage information between event signals, workflows, and tables. Explore the technical architecture, critical design considerations, and practical implementations through detailed explanations of interfaces, integrations, DSL, backfill workflows, and execution abstractions. Gain insights into why Netflix chose to build their own solution and how it has improved engineer productivity across their data ecosystem.
Syllabus
Introduction
Welcome
Landscape of Netflix
What is Maestro
Features
Why build our own solution
Scalability
User cases
Pipelines
Architecture
Akash
Critical Considerations
Interfaces
Integrations
Mastro DSL
Backfill workflow
Execution abstractions
Maestro execution
Taught by
AWS Events