METRO's Journey into Microservices: Experience and Lessons learned along the way
summary
METRO digitalizes a €3 billion delivery business across 750 stores by decomposing functions into 120 microservices and running them on kubernetes in production, reporting dozens of daily deployments and ~216 Java processes. The Betty delivery app and supporting CRM, driver and Salesforce tools use Java 8, React, Node and Go, REST synchonous calls and a custom deployment manager in Go. Pain points include cross-cluster service discovery, consolidated logging, slow integration tests and release gating for multi-country rollouts.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 1:44About METRO
- 2:10Scale of the business
- 5:28Customer Groups
- 9:34Digital Organization
- 14:24CustomerCentric Approach
- 14:43Customer Data Management
- 16:00Why Microservices
- 17:12Business Agility
- 17:26User Journey
- 18:42Statistics
- 19:33Mobilefirst
- 21:12DevOps
- 22:07Data Persistence
- 22:28Data Sharing
- 23:09Dont build your own
- 25:30Deploy Manager
- 26:54Integration Tests
- 27:11Challenges
- 36:47Booth
- 38:01Challenges of microservices
- 39:33Feature toggles
- 40:19User groups
- 41:39Future targets
- 42:59Why Rest
- 43:49Architecture
- 44:19Conclusion
related talks
51mCompetitive retail architecture with microservices
55mPatterns and practices for real-world event-driven microservices
38mPatterns and Practices for Real-world event-driven microservices
45mDesigning, building, testing and deploying microservices
51mMicroservices and Microfrontends: Scaling, Learning, Growing
01h 07mMicroservices from the Trenches. A Guide through the Architecture, Design, Code
35mSeven years with microservices: a critical review
38mFrom Monolith to Microservices at Zalando