Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Learn how Uber built a comprehensive deployment safety framework for managing thousands of Apache Flink jobs at scale in this 33-minute conference talk. Discover the engineering challenges of operating mission-critical streaming applications that power real-time decisioning, pricing, and operational workflows affecting millions of users. Explore Uber's systematic approach to deployment safety through their Flink-as-a-Service platform, which combines progressive rollouts, automated testing, and intelligent rollback mechanisms to maintain reliability without sacrificing development velocity. Understand the key components of deployment incrementality that limit blast radius through progressive rollouts, automation and CI/CD guardrails that ensure consistent code and configuration validation across environments, and comprehensive unit and end-to-end testing strategies that catch risky changes early through automated checks and traffic injection. Examine smart rollback systems with automated, metric-triggered responses that prevent widespread failures, and tenant-aware testing approaches that validate behavior across real workloads and production environments. Gain practical insights from Uber's experience scaling this framework to thousands of streaming jobs and learn how these principles can be applied to strengthen event-driven data platforms and build self-healing deployment pipelines for modern streaming systems.
