Azure Stream Analytics preview features

This article summarizes all the features currently in preview for Azure Stream Analytics. Using preview features in a production environment isn't recommended.

Public previews

The following features are in public preview. You can take advantage of these features today, but don't use them in your production environment.

Visual Studio Code for Azure Stream Analytics (Released May 2019)

Azure Stream Analytics jobs can be authored in Visual Studio Code. See our VS Code getting started tutorial.

Anomaly Detection

Azure Stream Analytics introduces new machine learning models with support for spike and dips detection in addition to bi-directional, slow positive, and slow negative trends detection. For more information, visit Anomaly detection in Azure Stream Analytics.

SQL Database reference data

Azure Stream Analytics supports Azure SQL Database as a source of input for reference data. You can use SQL Database as reference data for your Stream Analytics job in the Azure portal and in Visual Studio with Stream Analytics tools. For more information, visit, Use reference data from a SQL Database for an Azure Stream Analytics job.

Integration with Azure Machine Learning

You can scale Stream Analytics jobs with Machine Learning (ML) functions. To learn more about how you can use ML functions in your Stream Analytics job, visit Scale your Stream Analytics job with Azure Machine Learning functions. Check out a real-world scenario with Performing sentiment analysis by using Azure Stream Analytics and Azure Machine Learning.

JavaScript user-defined aggregate

Azure Stream Analytics supports user-defined aggregates (UDA) written in JavaScript, which enable you to implement complex stateful business logic. Learn how to use UDAs from the Azure Stream Analytics JavaScript user-defined aggregates documentation.

Live data testing in Visual Studio

Visual Studio tools for Azure Stream Analytics enhance the local testing feature that allows you to test you queries against live event streams from cloud sources such as Event Hub or IoT hub. Learn how to Test live data locally using Azure Stream Analytics tools for Visual Studio.

.NET user-defined functions on IoT Edge

With .NET standard user-defined functions, you can run .NET Standard code as part of your streaming pipeline. You can create simple C# classes or import full project and libraries. Full authoring and debugging experience is supported in Visual Studio. For more information, visit Develop .NET Standard user-defined functions for Azure Stream Analytics Edge jobs.

Other previews

The following features are also available in preview.

C# custom deserializer for Azure Stream Analytics on IoT Edge and Cloud (Announced May 2019)

Developers can implement custom deserializers in C# to deserialize events received by Azure Stream Analytics. Examples of formats that can be deserialized include Parquet, Protobuf, XML, or any binary format. Sign up for this preview here.

Parquet Output (Announced May 2019)

Parquet is a columnar format enabling efficient big data processing. By outputting data in Parquet format in a data lake, you can take advantage of Azure Stream Analytics to power large scale streaming ETL and run batch processing, train machine learning algorithms or run interactive queries on your historical data. Sign up for this preview here.

One-click integration with Event Hubs (Announced May 2019)

With this integration, you will now be able to visualize incoming data and start to write a Stream Analytics query with one click from the Event Hub portal. Once your query is ready, you will be able to productize it in few clicks and start to get real-time insights. This will significantly reduce the time and cost to develop real-time analytics solutions. Sign up for this preview here.

Support for Azure Stack (Announced May 2019)

This feature enabled on the Azure IoT Edge runtime, leverages custom Azure Stack features, such as native support for local inputs and outputs running on Azure Stack (for example Event Hubs, IoT Hub, Blob Storage). This new integration enables you to build hybrid architectures that can analyze your data close to where it is generated, lowering latency and maximizing insights. Sign up for this preview here.