A data transformation, normalization, and enrichment platform that handles all the details, so that you can focus on your business logic.

The Problem Space

Modern organizations collect and send streams of data to and from many partners, sensors, and other data sources. The one constant with these flows is change. It's a struggle to keep up with this flux and the operational complexity that comes with it.

How does DeltaFi help?

DeltaFi is a data-first platform that hosts your business logic and data model. It processes data in flight while providing data stewards unprecedented control and insight into every data flow.

Users can examine the data and accompanying metadata at each point in the system to understand how it was transformed.

When something goes wrong, DeltaFi allows you to see the details of the error, correct the error conditions, and rerun the affected data.

Engineers can focus on the core mechanics of data transformation and quickly create value through a configuration change or a few lines of simple code.

Features

Feature	Description
Cloud Ready	DeltaFi can be installed on any modern Kubernetes cluster via Helm Charts. Most DeltaFi components scale horizontally with minimal configuration tweaks.
Data Loss Protection	DeltaFi provides strong guarantees against data loss or orphaned data and metadata.
Metrics and Monitoring	Metrics are collected at the system, flow, Action, and data item level. Users can quickly identify and address bottlenecks and hot spots. Users can dig into centralized logs to uncover the root cause of any problems.
Tracking and Provenance	Every data item is tracked as it traverses the system. Users can see a complete graph of its journey.
Error Reporting	Users can customize the format and delivery of error reports that result from any processing errors.
Error Handling	If data processing fails because of a transient problem, users can resume processing at any time from the point that it originally failed, or replay data from the beginning.
Retention	Users can create policies dictating how long processed data should be retained.
Filtering	Users can terminate unwanted data mid-flow.
Reinjection	Data can be split into multiple children that can be injected into the front of the system for independent processing.