Detailed Course Outline
Phases
- Phase 1: The “Rusty” Rescue Mission
Teams will begin by deploying the existing "Legacy Monolith" application – a web API serving a frontend, to a single Azure App Service instance connected to an Azure SQL database. This establishes the baseline performance and first environment in the cloud. This challenge focuses on mastering Infrastructure-as-Code (Bicep) to get the initial legacy stack running live in the East US region.
- Phase 2: The Relational Jailbreak
Participants will act as data architects to analyze the rigid SQL schema and redesign it to support the flexible capabilities of Azure Cosmos DB. Teams must map relational tables, specifically the high-traffic shopping cart, into efficient JSON documents and migrate the data. This challenge emphasizes the critical shift from relational modeling to NoSQL thinking, requiring teams to define partition keys that will support future scale.
- Phase 3: Rewiring the Mainframe
Teams will dive into the application code to replace the legacy relational data access layer with the Azure Cosmos DB SDK. Participants must refactor the application logic to interact seamlessly with the new NoSQL backend, implementing efficient point reads and query operations. This challenge focuses on the practical application of the SDK and ensuring the application functions correctly with its new "brain" before any scaling occurs.
- Phase 4: Around the World in 80ms
To defeat the speed of light, teams will transform the single-region application into a globally distributed "Global Mesh" architecture. Participants must configure Cosmos DB to replicate data physically to Europe and Asia, while simultaneously deploying new App Service instances in those same regions behind Azure Front Door. This challenge requires implementing advanced local read patterns in the code, ensuring that a user in Tokyo automatically hits the Tokyo app instance and reads from the Tokyo database replica to achieve sub-50ms latency.
- Phase 5: The Influencer Apocalypse
Teams will face a simulated viral event where traffic spikes by 10,000%. Participants must demonstrate their system's resilience by configuring Autoscale on their Cosmos DB containers and App Service plans to handle the load without crashing. This challenge validates the entire architecture by testing whether the selected partition keys and scaling settings can sustain high throughput under extreme pressure while maintaining the low latency achieved in the previous phase.