How to Build an Event-Driven, Dynamically Reconfigurable Microservices Platform
summary
Describes a platform that combines Apache Ignite and Docker to build an in-memory, event-driven microservice grid where containers auto-register, auto-discover and form a software-defined network. Explains how the system achieves exactly-once processing, multi-datacenter high availability with automated replica promotion and live failover, dynamic reconfiguration and rolling upgrades with zero downtime. Gives implementation details: services running in hundreds of containers (demo ~500), automatic partitioning/rebalancing of primaries and replicas, and live upgrades replacing containers one at a time while processing continues.
AI-written summary · about
chapters (13)
- 0:00Event-driven intro
- 3:55Monolithic banking failures
- 7:56Distributed event design
- 12:33Domain-driven partitioning
- 16:29Event sourcing and state
- 20:08High-performance caching
- 23:37Transactional consistency
- 28:00Customer journey services
- 31:10Container orchestration
- 36:31Multi-datacenter failover
- 41:10Latency-aware deployments
- 43:42Open-source tools
- 44:04Q&A
related talks
01h 08mEvent Driven Architecture – Enabling Microservices
45mContinuously delivering containerized, highly available microservices
49mDeveloping microservices applications on the Kubernetes platform
51mCompetitive retail architecture with microservices
10mThe differences between microservices and containers
25mData-Driven Migration from Monoliths to Event-Driven Microservices
49mMicroservices: An unexpected Journey
56mHow to get a grip on your microservices system using a service-mesh