Lost in transaction Strategies to manage consistency in distributed systems
summary
Argues against distributed two‑phase commits and insists aggregates are the consistency boundaries, then shows a runnable Spring Boot example embedding a Camunda workflow engine to implement the Saga pattern with idempotent REST calls and compensating actions (demo uses a fake Stripe endpoint). Live demo exercises retries, time‑based waits, incident marking for stuck instances, and a UI; claims 50k/sec with relational backends and up to hundreds of k–1M/sec with an event/RAFT architecture.
AI-written summary · about
chapters (43)
- 0:00Introduction
- 0:36Domain term design
- 8:47Twophase commit
- 10:00Consistency
- 10:53Local asset
- 12:11Money transfer
- 13:17Order
- 14:24Distributed systems
- 17:27My favorite metaphor
- 19:06Network exception
- 20:08Payment exception
- 21:00Probability
- 21:26State handling
- 21:54Transparency
- 22:42Workflow Engines
- 23:33hexagonal architecture
- 24:44code example
- 30:59patterns
- 33:24scaling
- 33:45book
- 34:21Compensating transactions
- 36:36Proxy discussion
- 36:58Event chain
- 38:15Bad gut feeling
- 39:56Search the internet
- 40:21Order by risk
- 41:32Choreography
- 42:14Orchestration
- 43:06Compensation included
- 45:13Sources
- 45:45Graphical Models
- 46:06Sagas
- 47:16BPMN
- 47:29Living documentation
- 47:54Visualization
- 48:04Base DevOps
- 48:47Code
- 49:24Evaluation
- 49:46Tools
- 51:44Evaluate
- 54:02Questions
- 57:39What if this fails
- 59:18What if this is not synchronous
related talks
46mLost in transaction? Strategies to deal with (in-)consistency in distributed systems
36mLost In Transaction
30mData Consistency Patterns in Cloud Native Applications
01h 02mConsistency and Agreements in Distributed Systems
55mTackling transactions in Microservices applications
50mTransactional solutions for microservices
48mLost in transaction?
01h 38mBuilding Reliable Distributed Systems