Tackling Cloud Complexity - Recipes From Real World Projects
summary
Scalable event-driven systems get concrete recipes: a Node.js/Kubernetes stack using Kafka-style event stream inputs, measurable business metrics (connected devices, processing delay), and graceful shutdown to avoid data loss. Proven at half million devices, the team implements a thin producer router to split traffic across regions/subscriptions, uses distributed locking with Redis or partitioning to avoid race conditions, and supports canary/dedicated deployments; these approaches work in production.
AI-written summary · about
chapters (10)
- 0:00Introduction and speaker intro
- 8:54Data stream processing fundamentals
- 17:33Graceful shutdowns in cloud apps
- 21:28Distributed locking for consistency
- 25:02Partitioning strategies explained
- 31:10Scaling beyond event stream limits
- 36:54Input and output routing techniques
- 40:30Canary deployments and testing
- 43:03Universal recipes for cloud systems
- 45:55Q&A and closing remarks
related talks
29mFollowing the Data: A Deeper Look into Distributed Applications
28mCloud Agnostic Design Patterns and Tips
54mReacting to an Event-Driven World
24mShaping Tomorrow's Technology: Navigating Cloud-Native, Serverless, and Polyglot Programming
25mOperating Cloud-Native Kafka
26mTen Best Practices for Event Streaming Success
46mPracticalities of Productionizing Distributed Systems
48mCloud Native Event Streaming with Kafka & Open Standards