Distributed Workflows for Microservices-Style Applications [I]
summary
A lightweight workflow library implements cross‑service business flows using NirmataOSS Workflow (Java + ZooKeeper + Curator) and models task relationships as task dependency as DAG. It schedules task nodes to distributed executors, supports automatic retries and resume after node or ZooKeeper failures, and demonstrates updating container images across many microservices on Kubernetes where independent tasks run concurrently and dependent tasks wait, showing the pattern works in production. Mentions Netflix Conductor and AWS SWF as alternatives.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:10Microservice Architecture
- 1:25About Me
- 2:08Outline
- 2:35Microservices Architecture
- 3:02What is Workflow
- 4:08Microservices
- 5:20Features
- 6:05Workflow Manager
- 7:13Task Execution
- 10:38Example Code
- 11:05Demo
- 15:57Other great solutions
- 16:55Questions
related talks
22mManaging Distributed Workflows at Scale
30mMessage-driven microservices orchestration
31mBuilding Asynchronous SOA for Modern Apps
38mDistributed Workflows with Flowy
39mWorkflows, a new abstraction for distributed systems
51mChoreography vs Orchestration in serverless microservices
39mEnabling Microservices at Orbitz
36mOnyx: Distributed Workflows for Dynamic Systems
