Choreography vs Orchestration in serverless microservices
summary
Compares event-driven choreography with a central orchestrator and shows a concrete rewrite: an event-heavy image app called Pika Daily is converted to HTTP-driven orchestration using Google Cloud Workflows. Demonstrates removing parsing for three different event formats, cutting glue code, centralizing retries/timeouts and getting a visual flowchart and per-step failures, while trading away parallelism and adding YAML-debugging and a new service to learn. Includes live demo of uploads, metadata extraction, thumbnails and collages succeeding under the workflow.
AI-written summary · about
chapters (11)
- 0:00Introduction and welcome
- 1:05Talk overview and context
- 3:31Choreography vs orchestration basics
- 8:11Event-driven choreography pros and cons
- 13:26Orchestration approach explained
- 20:07Tools and services landscape
- 23:35Google Cloud Workflows deep dive
- 33:01Case study introduction
- 37:35Rewriting to orchestration
- 40:46Lessons learned from rewrite
- 43:55Resources and closing
related talks
51mChoreography vs Orchestration in serverless microservices
50mChoreography vs Orchestration in serverless microservices
46mChoreography vs Orchestration in serverless microservices
46mChoreography vs Orchestration in serverless microservices
50mChoreography vs Orchestration in serverless microservices
01h 03mChoreography vs Orchestration in serverless microservices
32mChoreography vs Orchestration in serverless microservices
38mChoreography vs Orchestration in serverless microservices
