Last month, Datadog traveled west to Seattle for Microsoft Build. Following a packed three days of keynotes, demos, breakouts, and meetings, we’ve collected the key announcements most relevant to the observability and security problems Datadog solves for Azure customers across on-prem, hybrid, and multi-cloud environments.

New large language models within Azure AI

As expected, new large language models (LLMs) and feature updates were announced, including the general availability of OpenAI’s latest frontier model, GPT-4o, via Azure OpenAI Service as an API. Microsoft also announced updates to its own small language models in Azure AI, with the general availability of its Phi-3 open models: Phi-3-small, Phi-3-mini, and Phi-3-medium. Through Datadog’s existing Azure integration, Datadog users can already monitor performance, usage, and costs of their Azure OpenAI API.

Beyond new frontier models, Microsoft announced the general availability of Azure AI Studio, a new platform for developing and embedding generative AI into application architectures. With the cost and complexity implications of generative AI, Datadog has been laser-focused on helping organizations deliver on these generative AI experiences in their production applications via our observability and security offerings. Datadog LLM Observability, now generally available, enables companies to continuously monitor and troubleshoot the AI-specific components of their stack in production. For best practices, see how Datadog monitors machine learning (ML) models in production and managed ML platforms.

Bringing AI capabilities to the database

In line with innovations across large and small language models, there were several updates focused on helping customers modernize their data estate to deliver better AI experiences.

A couple announcements in this area that stood out were the general availability of the Azure Database for PostgreSQL Azure AI extension and the preview of built-in vector database capabilities for Azure CosmosDB for noSQL. These managed Azure databases are enabling organizations to deliver low-latency, high-availability AI application experiences across relational and non-relational workloads. Datadog Database Monitoring is a critical product for Azure customers who need deep visibility into database performance across all on-prem and managed Azure database services. Database Monitoring supports self-hosted and managed cloud versions of PostgreSQL, MySQL, Oracle, and SQL Server.

And with Datadog Azure Monitoring, customers are already able to monitor performance and set up high-signal alerts for Azure CosmosDB and other critical managed database services in Azure.

Increasing developer productivity

To support developers in building new AI apps, Microsoft announced several new features across Azure App Service, Azure Kubernetes Service, Azure Container Apps, and Azure Functions. One update for Azure App Service was the preview of Sidecar patterns, which enables customers to add extra monitoring and logging capabilities to applications without having to make changes in the code. Before Build, we published a step-by-step guide with Microsoft explaining how to use Sidecars to integrate Datadog with your .NET applications hosted on Linux App Service.

As an Azure Native ISV Service, Datadog has continued to improve monitoring capabilities of these services, with customers today taking advantage of 40+ Datadog-generated metrics and dozens of new tags for their Azure services. For example, this year, we streamlined Azure container monitoring with our Datadog Azure Kubernetes Service cluster extension.

Datadog was featured in two Build breakout sessions relevant to developer productivity: “Accelerate DevOps & Incident Management workflows with Microsoft Teams” and “Using AI with App Service to deploy differentiated web apps and APIs”. Datadog was also mentioned during the day one keynote along with other companies solving problems for DevOps professionals.

See you next year

This was just a short summary of the new features and updates that excited us at Microsoft Build this year. For more announcements from Build, visit the Microsoft Build 2024 Book of News. And be sure to look out for more great content from us about how you can build on Azure. We hope to see you at Build next year.