Choreography vs Orchestration in serverless microservices
summary
Compares Pub/Sub event choreography versus centralized Cloud Workflows orchestration by rebuilding a photo‑sharing app: original pipeline uses Cloud Storage triggers, Cloud Functions, Cloud Run, Vision API and Firestore; rewrite moves logic into Workflows YAML that sequences HTTP calls, retries, timers and callbacks. Result: orchestration reduces glue code, simplifies error tracing and removes extra Pub/Sub setup, but YAML limits complex logic, local testing is hard and some microservices disappear; recommends a hybrid: orchestrate tight sequences, eventing for loose integrations.
AI-written summary · about
chapters (17)
- 0:00Start
- 5:23How Error Reporting or Retry and Timeout Logic Works
- 8:45Pros and Cons
- 13:22Orchestration
- 19:25Compensation Steps
- 23:28Have an Error on some Step and We Need To Show It on the Ui Do We Better Notify Ui Directly or via Orchestration
- 29:52Is It Ideal and Feasible for an Orchestrator To Be Simple Enough that It Can Demand a Hundred Percent Test Coverage
- 32:41Cloud Events
- 32:59Open Api Specification
- 35:26Define Your Workflow
- 39:37Other Useful Features
- 40:09Iteration
- 40:22Callbacks
- 41:15Visualization of the Orchestration
- 49:03Garbage Collection
- 56:18Google Cloud Workflows
- 57:19Lessons Learned
related talks
46mChoreography vs Orchestration in serverless microservices
51mChoreography vs Orchestration in serverless microservices
50mChoreography vs Orchestration in serverless microservices
46mChoreography vs Orchestration in serverless microservices
49mChoreography vs Orchestration in serverless microservices
32mChoreography vs Orchestration in serverless microservices
50mChoreography vs Orchestration in serverless microservices
50mChoreography vs Orchestration in Serverless Microservices