Data Consistency in Microservices Architecture
summary
Describes five concrete approaches to keep entities consistent across microservices: Sagas with compensating operations for recoverable failures, change data capture that writes once then streams updates, and event-first architectures using Kafka plus tooling like Kafka Connect to stream DB changes into topics. Cites production experience (eBay Classifieds, payments and lending systems), tradeoffs (2PC/XA latency and coordinator availability), when teams merge services back, and when each approach succeeds or adds operational complexity.
AI-written summary · about
chapters (20)
- 0:00Intro
- 0:23Why Data Consistency Matters
- 1:46Why Microservices Architecture
- 2:21Data Consistency Patterns
- 3:35Compensating Operations
- 4:23Reconciliation
- 5:28End of Day Procedures
- 6:16How we can reconcile
- 7:27Complex reconciliation
- 7:58Application Aware Login
- 10:02Standard Solution
- 11:05Seed Guarantee
- 12:11Change Data Capture
- 14:36Techniques and Solutions
- 15:45Challenges
- 17:02EvenDriven Architecture
- 23:32My Choice
- 25:04Consistency Challenges
- 25:42Designing Data Intensive Applications
- 26:28Questions
related talks
46mMicroservices architecture and data consistency
39mManaging Consistency, State and Identity in Distributed Microservices
47mMicroservices Architecture and Data Consistency: So How Do We Do It for Real?
51mPatterns & Practices of Implementing Event-driven MicroSvc...
01h 11mEmbracing Messaging and Eventual Consistency in your Microservices Solutions
30mData Consistency Patterns in Cloud Native Applications
54mTackling transactions in a Microservices application
48mPG Saga: Coordinating Dependent Data Changes Across Multiple Services