Integrating Azure AI and Azure Kubernetes Service to build intelligent apps
summary
Shows a working eShop example that adds a chat microservice backed by Azure OpenAI Service using the Semantic Kernel, fronted by a Vue partial and hidden behind Azure App Configuration feature flags. Deployment runs on AKS with the new Istio add-on on AKS (managed control plane), Prometheus/Grafana, Flagger for canary/dark launches and KEDA autoscaling tied to Azure Load Testing. Demo walks through GitHub Actions, traffic splitting, a 250-user load test (ramped in 60s) where KEDA scales pods and P90 latency drops, proving the pieces integrate end to end.
AI-written summary · about
chapters (19)
- 0:00Introduction to AI apps
- 1:51Responsible AI and principles
- 3:43Service mesh fundamentals
- 5:51Istio add-on announced
- 7:42Istio operator benefits
- 10:47AI application architecture
- 12:54Deploying the AI bot
- 15:00Bot demo and semantic kernel
- 17:53Feature flag configuration
- 19:27Istio traffic management
- 22:34Automated canary deployments
- 27:00Keda auto scaling demo
- 30:42Scaling and load impact
- 32:24AKS ecosystem overview
- 34:21AKS co-pilot announced
- 36:25Co-pilot operations demo
- 38:27Kubectl AI plugin
- 40:43Co-pilot diagnostics
- 43:22Key takeaways and QA
related talks
01h 01mDevOps with Azure Kubernetes Service
43mUsing AI with App Service to deploy differentiated web apps and APIs
11mThe Future of Cloud Native and AI
57mServerless on AKS (Azure Kubernetes Service)
51mUsing AI to Deploy Java apps to Kubernetes
01h 01mKubernetes, Azure and .NET - What's it all about?
47mModern continous delivery on Azure Kubernetes Service
40mDeploying Micro Frontends to a Kubernetes Cluster