Service Bus is Azure’s fully managed enterprise message broker. It supports Queues (point-to-point) and Topics/Subscriptions (pub/sub). Choosing the right pattern prevents architectural headaches. Queues vs Topics Dead-Letter Queue Messages that fail processing N times go to a special DLQ for inspection and replay. Key Takeaways Use **Sessions** for ordered processing (FIFO for a specific session […]
Read more →Search Results for: events
Managing Terraform State in Azure
Terraform’s `.tfstate` file is gold. If it’s lost or corrupted, Terraform cannot track what resources exist. You **must** store state remotely with locking. Azure Backend Configuration State Locking Azure Blob’s native lease mechanism prevents concurrent writes. Key Takeaways Never commit `.tfstate` to Git. Enable **soft delete** on the storage account to recover corrupted state. Use […]
Read more →Azure Container Apps: KEDA and Dapr Managed
Azure Container Apps (ACA) is the “Serverless Containers” offering we’ve been waiting for. Built on Kubernetes but hiding the cluster, it integrates **KEDA** (for event-driven scaling) and **Dapr** (for building blocks) natively. Scaling to Zero Unlike App Service, ACA can scale to 0 replicas when no events are processing, saving money. Dapr Integration Enable Dapr […]
Read more →Azure Web PubSub: Real-time WebSockets
SignalR is great, but sometimes you want raw WebSockets or support for other languages (Python, Java clients). **Azure Web PubSub** is a managed WebSocket service that supports native WebSocket clients and the PubSub subprotocol. Architecture The service handles the massive concurrent connections. Your server only handles events (Connect, Message) via Webhooks. Key Takeaways Supports standard […]
Read more →C# 9 Source Generators: Removing Reflection
Reflection is slow. It happens at runtime, bypasses type safety, and prevents trimming. Source Generators solve this by generating code at compile time. In this guide, we build a generator that automatically implements a `MapTo` method for DTOs, replacing AutoMapper. The Goal The Generator Logic Key Takeaways Source Generators enable **Zero-Overhead abstractions**. They are essential […]
Read more →Angular State Management: NgRx vs Akita
State management in Angular is often over-engineered. Do you really need the full Redux pattern (NgRx) with its boilerplate Actions, Reducers, Effects, and Selectors? Or is Akita’s OO-approach better? NgRx: The Strict Pure Approach NgRx is verbose but predictable. It shines in large teams where strict enforcement of “One Way Data Flow” prevents spaghetti code. […]
Read more →