This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.
Case Study
How Fabric Helps Envisor Deliver Continuous Cloud Cost Analytics at Enterprise Scale
Industry:
Technology
Focus:
FinOps
Envisor rearchitected its cloud cost management (CCM) platform by moving from a traditional Azure SQL Server data warehouse to Microsoft Fabric. Built on OneLake, the new medallion architecture enables near-real-time analytics, dramatically improved performance and reliability, and reduced operational and capacity costs.
Challenge
Envisor helps organizations monitor and control cloud spending by delivering timely insights across cloud and IT environments. Its AnyScope™ platform ingests and processes extremely large, highly granular cloud cost datasets, including CUR and FOCUS files.
Envisor’s original back-end architecture reflected best practices at the time and was designed to support the then-nascent FinOps/cloud cost management (CCM) landscape. Built on a traditional Azure SQL Server data warehouse with Power BI as the analytics layer, the platform operated on a nightly batch processing model.
As FinOps analytics sophistication accelerated, customers sought visibility across additional technology scopes such as SaaS, AI and on-prem. They also wanted to integrate FinOps insights directly into their broader BI environments, putting cost intelligence alongside operational and financial data.
To meet these evolving demands, the Envisor platform team knew it would need a more robust and scalable architecture, one capable of handling growing data volumes, diverse sources systems and increased demand for near-real-time cost intelligence.
Large-scale rebuilds & long reprocessing times
- Schema or logic changes often required full data warehouse rebuilds.
- Even modest changes triggered long reprocessing cycles.
- Rebuilds routinely took ~24 hours.
- Extended processing times for enterprise customer datasets were threatening SLAs
Strain due to growing Power BI semantic models
- Semantic models were update nightly and continued to grow with data volumes.
- Larger models required more compute time and longer refresh windows.
- DirectQuery performance constraints made it unsuitable for reducing model size.
Rising compute and operating costs
- Long processing times necessitated sustained compute usage.
- Database rebuilds required scaling compute to much higher tiers.
Solution
Envisor evaluated unified data platform options capable of supporting enterprise-scale workloads. A key requirement was the ability to move away from full database rebuilds and batch processing, while preserving access to detailed raw cost data.
Envisor selected Fabric with OneLake for its Microsoft-native integration and dramatically lower cost and complexity compared to Databricks. The medallion architecture fundamentally changed how data was ingested, processed and managed in the Envisor platform:
- Data is processed incrementally, rather than rebuilt end-to-end.
- Transformations are layered, with clear separation between Bronze (raw), Silver (refined) and Gold (curated) data.
- Semantic models moved to Direct Lake, preserving fast performance and cutting compute costs by eliminating refresh overhead.
- Original source data is preserved permanently in OneLake’s Bronze layer.
- Transformations are layered, with clear separation between raw, refined and curated data.
Benefits
With Fabric, Envisor turned its platform into a powerhouse. It unlocked the large-scale enterprise power, speed and reliability the platform needed to go from great to exceptional.
Dramatically reduced processing times
- Enterprise-scale datasets now process in under an hour instead of nearly 24 hours during month-end.
- Power BI semantic model data refreshes are no longer needed, improving access to fresh data.
Significantly lower operating costs
- Overall compute costs reduced by 50% or more compared to the legacy architecture
- Direct Lake allowed some clients to move to a small capacity, resulting in savings as high as 75%.
- Enterprise-scale workloads can be supported at lower cost, even when accounting for Power BI capacity.
Faster development and iteration
- Schema and logic changes no longer require full rebuilds.
- New data elements and transformations can be added incrementally.
Expanded analytics and AI readiness
- Pristine original data is always available to ML and AI models.
- Enables advanced analytics and AI without re-ingestions.
While we anticipated performance gains from Fabric, the results far exceeded our expectations. The speed alone is compelling, and when combined with reduced operational overhead and greater flexibility, the ROI is undeniable.
– Keith Knowles, Managing Director, Envisor
About Envisor
Envisor is a leading provider of enterprise cloud and tech spend management solutions. Built on an open, Microsoft Fabric-native platform and FOCUS-compliant, Envisor accommodates the reporting flexibility, cloud security needs, and data integration capabilities required for Run-stage FinOps. Envisor is an active contributor to the FOCUS specification and a dedicated partner of the FinOps Foundation. Envisor is headquartered in Burlingame, CA and is a wholly owned subsidiary of Senturus, a national data and analytics consulting firm.