Adventures of building a multi-tenant PaaS on Microsoft Azure
summary
Builds a multi-tenant PaaS on Azure using Functions, App Services and AKS with KEDA on AKS to handle app and cluster autoscaling including scale-to-zero and roughly 20 scaler types observed in practice. Uses sharding with a lightweight shard manager and elastic pools per-tenant to control costs and per-shard caps. Routes third-party webhooks through API Management for webhook proxy to preserve security and retries. Adds health endpoints, correlated telemetry, automated alerts and deployment rings. Platform ships in under a year but already requires service migrations.
AI-written summary · about
chapters (11)
- 0:00Introduction to multi-tenant PaaS
- 3:15Building multi-tenant platforms on Azure
- 4:43Basics of scaling applications
- 15:53Good, bad and ugly of scaling
- 23:50Multi-tenancy and data isolation
- 35:22Monitoring and telemetry strategies
- 45:23Health checks and alerts importance
- 51:30Webhooks and event-driven architecture
- 1:03:01Embracing change and automation
- 1:06:41Cloud service lifecycle and portability
- 1:17:31Conclusion and key takeaways
related talks
46mModernizing your applications with containers and serverless
50mEvent-driven autoscaling on Kubernetes with KEDA and Azure Functions
41mEvent-driven autoscaling on Kubernetes with KEDA and Azure Functions
45mLeverage AKS for your enterprise platform: H&M’s journey
35mBuilding Apps with Containers, Functions and Managed Services
29mNavigating the Multitenancy Maze: Avoiding Kubernetes Antipattern Pitfalls
39mLessons Learned Migrating an Existing Product to a Multi Tenant Cloud Native Environment
40mScaling Startup for Success: Building modern Architectures and Developer Acceleration