Azure Weekly

Issue #513 - 4th May 2025

We're in the pre-BUILD 2025 lull, but there are still a large number of announcements and in depth articles. To kick off, One year of Phi: Small language models making big leaps in AI, and Azure Compute Fleet is Generally Available, and Azure Functions - Learnings from executing PowerShell from C#.

In Analytics, there are two good Fabric articles from the community: Automate testing Microsoft Fabric Data Pipelines with YAML Pipelines, and Microsoft Fabric RTI: Azure Event Grid Namespace MQTT support.

In Cloud Native, there are two interesting posts about networking advancements: What's New in the World of eBPF from Azure Container Networking, and Introducing Layer 7 Network Policies with Advanced Container Networking Services for AKS Clusters, and a detailed guide to ArgoCD integration with Private AKS Cluster.

Finally, Barry Smart has written a great in-depth 3 part series on DuckDB - where he makes the case that if you deal with data of the ~1 billion rows magnitude, you might be better served using the high-performance, in-process analytics engine, rather than a full-blown cloud data platform. Part 1 is DuckDB: the Rise of In-Process Analytics and Data Singularity, Part 2 is DuckDB in Depth: How It Works and What Makes It Fast, Part 3 is DuckDB in Practice: Enterprise Integration and Architectural Patterns, and the code samples are available on GitHub. It's about a 30 minute read, but essential if you do any data engineering work, and want to improve the speed of your inner-dev-loop.