Microsoft introduces Drasi: open source system for real-time event processing and automation

Microsoft’s Azure Incubations team introduced Drasi, an open source system that simplifies the detection of critical events in complex infrastructures. Drasi provides real-time monitoring and automated responses, eliminating the need for manual event handling. With flexible components and integrations, it optimizes change detection across different data sources. Available under the Apache 2.0 license, Drasi provides developers with a customizable solution for managing event-driven systems.

Drasi addresses the challenges of scaling event-driven systems by offering an integrated solution to monitor changes and automate appropriate responses. By providing a unified framework that reduces complexity and ensures timely responses, developers eliminate the need to manually create event handling mechanisms.

At its core, Drasi works with three main components:

1. Sources: These monitor various data points including application logs, database changes and system metrics. Sources track changes in real time and capture data without copying it to central repositories or data lakes.
2. Continuous queries: Instead of relying on traditional point-in-time queries, Drasi uses continuous queries that evaluate data as it arrives. Written in Cypher Query Language, these queries integrate information from multiple sources, constantly updating results without manual intervention.
3. Reactions: When changes meet the criteria defined in a continuous query, Drasi triggers automated responses. This can include tasks such as sending alerts, updating systems, or taking remedial actions based on specific business needs.
   

Drasi’s architecture (Source: Microsoft Blog)

Drasi is flexible and extensible, allowing users to integrate custom sources and reactions based on their individual needs. Pre-built integrations with platforms such as PostgreSQL, Microsoft Dataverse and Azure Event Grid are also available.

This system provides a more efficient alternative to traditional data processing methods based on constant polling, polling or batch processing. By eliminating delays and processing overhead, Drasi ensures organizations can respond to critical events as they occur, helping to avoid missed opportunities and risks.

Allen Jones, CTO at Microsoft, shared the following note:

Drasi is the project that I have been leading in the Azure Incubations team for several years. It’s a product that I could have used so many times in my career and it’s exciting to bring it out to the world, see what the community thinks about it and understand how they’ll use it.

The feedback from the community was positive. For example, Tejaswi Redkar wrote:

Drasi seems relevant to most operations domains where data -> insights -> actions are the norm, but connecting the three to automate is still a challenge.

Drasi can be used via GitHub and is licensed under Apache 2.0.