Building Stateful Clustered Microservices with Java, Actors, and Kubernetes
summary
Java microservice uses Akka actors plus Kubernetes to run a stateful clustered shopping-cart demo that actually works: a Maven build produces a Docker image, kube deployment runs 3 pods, Akka cluster bootstrap queries Kubernetes to find peers, and a sharding layer keeps 15 shards distributing entity actors (one actor per cart). Live demo shows pods killed and auto-recreated, shards rebalanced, and manual scale to five pods with actors migrating; no database persistence in this example, but full clustering and rebalancing operate correctly.
AI-written summary · about
chapters (31)
- 0:00Introduction
- 0:53What is this talk about
- 2:58The Actor Model
- 9:23Kubernetes
- 11:24Demo
- 11:55Pods
- 12:57Visualization
- 13:49Crop Circle
- 15:01Actors
- 16:59Microservices
- 17:21Resiliency
- 19:10Kubernetes Deployment
- 21:40Reactive Systems
- 22:58The Project
- 24:14The Deployment File
- 26:20The Main Project
- 34:45The Entity Actor
- 38:20Visualizations
- 39:18Playing with code
- 40:40Cluster Actors
- 42:12JSON Tree
- 42:56Code
- 47:07Cluster Singletons
- 47:48Core Persistence
- 48:38Event Persistence
- 49:23Kubernetes Relationship
- 49:42Projects
- 50:51Comic Book
- 51:39Shard vs Pod
- 53:20The mechanics of distributed systems
- 54:23Is it possible to run this without Kubernetes
related talks
54mJava based microservices, containers, Kubernetes - how to
45mReactive Microservices with Akka and Docker
49mManaging an Akka Cluster on Kubernetes
51mJava-based microservices, containers, Kubernetes - how to
46mReactive Microservices with Akka and Docker
46mModern DevOps with Kubernetes
48mThe three paths of Scaling with Akka
38mManaging an Akka Cluster on Kubernetes