Scaling microservices with Message queues, .NET and Kubernetes
summary
Shows a .NET Core web front end, SignalR updates and a MongoDB-generated group ID feeding jobs into RabbitMQ, with worker pods that host two containers (console processor plus a metrics web endpoint). Kubernetes Horizontal Pod Autoscaler (HPA) runs between min 10 and max 20 replicas and automatically doubles workers under load; during the demo the HPA scales from 10 to 20, Grafana/Prometheus show queue depth fall (from ~443 toward ~137) and the jobs are processed.
AI-written summary · about
chapters (38)
- 0:00Introduction
- 2:35Why containers
- 4:31Demo
- 4:58Sam Newman
- 6:18CN
- 7:23Learn back
- 7:55Learn simple things
- 10:00Planning for launch
- 10:47Performance tests
- 12:00Too little too late
- 12:30Game over
- 13:19Decoupling scaling
- 13:45Decoupling failures
- 14:09Message queues
- 14:46Automatic scaling
- 15:29Design skills
- 16:14Containers
- 16:31Deployment to Kubernetes
- 17:30What are nodes
- 18:33Best Tetris player
- 19:04Users
- 19:14Demo Time
- 19:44Apologies
- 20:00Ticket
- 20:54Thank you
- 21:40Lets have a look
- 23:47Thank you sir
- 24:11HPA
- 25:03Application
- 27:03Metrics
- 27:41Check HPA
- 29:54Check IQ
- 30:15Check Rest
- 30:47Proxy Data
- 31:33Prometheus Dashboard
- 33:14Start a spike
- 34:33Questions
- 35:01Cloud hosting
related talks
42mScaling microservices with Message queues, .NET and Kubernetes
01h 04mNavigating microservices with .NET Core
35mKEDA - Real Time And Serverless Scaling In Kubernetes
34mAutoscaling Made Easy with Kubernetes Event-Driven Autoscaling (KEDA)
53mNext Generation Microservices
01h 00mEvent-driven computing with Kubernetes
46mModern DevOps with Kubernetes
34mScale your microservices by MicroProfile Metrics on Kubernetes