Scaling Kubernetes-Native Systems with Event-Driven Patterns on Azure
summary
Claim: an Azure pipeline using Event Grid → Service Bus → Azure Functions → GPU AKS pods scores transactions in under 2 seconds with ~95% accuracy, cuts false positives by 15% and sustains 100% uptime during peak loads handling millions of events per second. Implementation details include ordering via Service Bus sessions, enrichment functions, ML inference on GPU-backed AKS pods, idempotent handlers, semantic event versioning and correlation IDs for full observability.
AI-written summary · about
chapters (10)
- 0:00Introduction and Speaker Background
- 0:38The Importance of Event-Driven Kubernetes Systems
- 1:22Understanding Cloud Native and Event-Driven Systems
- 1:52Kubernetes and Event-Driven Architecture
- 4:13Microsoft Azure's Event-Driven Toolkit
- 5:28Deep Dive into Event Grid and Service Bus
- 7:29Azure Functions and Architectural Patterns
- 10:01Real-Time Fraud Detection Case Study
- 14:18Lessons Learned and Industry Applications
- 16:22Conclusion and Next Steps
related talks
47mModern continous delivery on Azure Kubernetes Service
56mReal-Time, Distributed Applications w/ Akka.NET, Kubernetes, .NET Core, and Azure Kubernetes Service
16mCracking the code: AI workloads at scale on AKS
38mHow Event Driven Autoscaling in Kubernetes Can Combat Climate Change
47mBuild scalable cloud-native apps with AKS and Azure Cosmos DB
20mHow are You Utilizing GPUs? Best Practices on Managing GPUs in Azure
47mIntegrating Azure AI and Azure Kubernetes Service to build intelligent apps
49mIntroducing AKS Automatic: The easiest managed Kubernetes experience