This blog is a preview of my session at Sirus Connects on October 9, 2026, where I’ll share practical lessons on building a reliable Microsoft Fabric data platform, covering capacity management, maintainable architecture, embedded analytics and data security.
Speaker:
"From sensitive data and limited capacity to embedded analytics: practical lessons from a year of working with Microsoft Fabric."
How do you build a modern data platform when budget, capacity and development time are not unlimited? And what happens when your data comes from different sources, includes sensitive information and ultimately needs to be made available in an external application?
Over the past year, I have been tackling these questions while implementing a reporting platform in Microsoft Fabric. Along the way, I discovered that a successful Fabric implementation is about more than choosing the right technology. How you design your architecture and use the capacity available to you can make just as much of a difference.
Building a maintainable data platform
In my session, I’ll take you through some of the main technical challenges I encountered along the way. Rather than focusing on one specific business case, I’ll share lessons and patterns that can be applied to other Fabric projects as well.
How do you build a maintainable data platform when you need to bring together data from different sources? How should you structure your pipelines and notebooks? And how do you make sure processes such as incremental loading, reusable transformations and historisation remain easy to maintain as your platform evolves?
These are choices that may seem small at first, but they can have a major impact on how manageable and scalable your data platform becomes over time.
Getting more out of your Fabric capacity
Another aspect that sometimes gets the attention it deserves only later in a Fabric project is capacity management.
A Fabric capacity has its limits, especially when you’re working with a restricted budget. I’ll show how the configuration of notebooks and Spark pools can affect capacity consumption and which choices can help you get more out of the capacity you already have.
We’ll also look at bursting, smoothing and throttling, and what they mean in practice. Before simply moving to a higher capacity, there are often several ways to optimise your existing setup first.
The goal is not just to keep costs under control. It’s to understand how your workloads behave so you can make informed decisions about performance, capacity and cost.
Secure embedded analytics without unnecessary licences
Finally, I’ll dive into embedded analytics and data security.
Integrating Power BI reports into an external application does not necessarily mean that every end user needs their own Power BI licence. I’ll discuss how Power BI Embedded and a service principal can be used to make reports securely available to external users, including the use of row-level security.
Sensitive data from on-premises environments brings its own set of challenges. How do you bring that data into Fabric securely while keeping your data platform scalable and manageable?
It’s an important part of the architecture that needs to be considered from the start, rather than added as an afterthought.
Practical lessons from a year of working with Fabric
Don’t expect a theoretical overview of Microsoft Fabric. Instead, I’ll share practical insights from a year of building, testing, optimising and adjusting a real data platform.
I’ll walk you through the architectural choices that worked well, the configurations that had a surprisingly big impact and the things I would approach differently with what I know today.
The aim is simple: to give you practical lessons you can apply to your own Fabric projects and help you make better-informed decisions about architecture, capacity and security.
Come to my session at Sirus Connects and discover how to get more out of Microsoft Fabric without letting your architecture, capacity or budget get out of hand.
Want to follow this live?
Join this session at Sirus Connects on October 9, 2026, from 13h30 to 14h15. Register and secure your spot.
Can’t make it? Reach out to us, we are happy to share our experience and discuss how to build a scalable, secure and cost-efficient data platform in Microsoft Fabric.
