Knative Serverless to architect decomposed components
summary
Demo decomposes a monolith into three microservices (account, address, person) deployed on OpenShift using Knative Serving and Eventing, proving HTTP-backed services scale from zero and handle real events. Database changes use Debezium outbox into Kafka, then Debezium -> Kafka feeds Knative with a Knative Source-to-Sink that posts to services via a Channel then Broker/Trigger filters. Services run as Quarkus containers with fast startup and small images, and the live demo shows the pipeline actually updating the account read model.
AI-written summary · about
chapters (35)
- 0:00Introduction
- 0:58Kubernetes complexity
- 2:40Serverless
- 3:18Knative
- 5:36Revisions
- 6:34Autoscaling
- 6:56Scaling
- 9:04Events driven architectures
- 10:48Cloud events
- 11:35Native Events
- 11:59Source to Sync
- 13:13Channels and Subscriptions
- 14:01Brokers and triggers
- 15:07Source binding
- 15:42Channels and Brokers
- 17:11Knative Eventing and Knative Serving
- 18:22Screen Extension
- 19:16Demo
- 22:28Klin
- 23:53Openshift
- 25:26people API
- 25:52graphql
- 26:30decomposing monolith
- 29:16Kafka Connector
- 29:37Script
- 31:13Openshift microservices
- 32:23Openshift Kubernetes
- 34:02Accounts
- 34:25Address Data
- 34:40Native Channel
- 38:15Database
- 40:38Person Microservice
- 41:53Personal Object
- 42:07quercus
- 43:19conclusion
related talks
23mServerless with Knative
52mServerless with Knative
49mServerless Java on Knative with Quarkus
45mKubernetes Serverless with Knative
26mKnative makes Developers Incredible on Serverless
35mStop Calling Knative Serverless!
48mEvent-driven serverless architectures using Knative and Cloud Run.
54mHomemade Serverless: Making it from Scratch