Case Study: Apache Superset keeps dashboards unbroken with Preset and Cube

Apache Superset keeps 3,073-word dashboards resilient with Preset

The customer, Apache Superset, a popular open-source data visualization tool, faced a significant challenge: its dashboards were frequently broken by inevitable upstream changes in the underlying data pipeline, such as renamed columns, removed tables, or entire data source migrations. These disruptions caused errors and interruptions for users, leading to unproductive work hours and unreliable analytics. To address this fragility, the vendor, Preset, which offers a managed cloud service for Superset, was used in conjunction with Cube's headless BI platform to create a semantic layer.

The solution implemented by Preset and Cube involved inserting a semantic layer between Superset and the raw data sources using Cube. This layer acted as a buffer, defining consistent metrics and dimensions. When upstream changes occurred, the data model in Cube was updated instead of the Superset dashboards, keeping them intact. The case study demonstrated how this approach successfully prevented dashboard failures when columns were removed, tables were altered, or data sources were migrated. The result was a robust and reliable Superset environment where dashboards remained functional despite upstream volatility, and Preset's integration with Cube's alerting system provided early warnings of potential issues.

Apache Superset