Microsoft Fabric Architecture Podcast

Matthias Falland

Architecture decisions for Microsoft Fabric. Anonymized real customer scenarios, cost realism, counter-arguments included. Weekly episodes aligned with Fabric Friday recordings.

  1. 3d ago

    A Label Is Not a Lock

    A Label Is Not a Lock Episode 43 • 2026-10-30 Duration: 9:02 A sensitivity label says Highly Confidential in the header, but inside Fabric it never decides who opens the item. This episode traces where labels actually control access and where the appearance of a lock is just a sign. What we discuss The pattern we keep seeing in the field How it actually works underneath the abstraction Where the obvious answer breaks A real Reddit/Microsoft Q&A question unpacked F-SKU realism — what this actually costs When the rejected approach is actually right The concrete recommended architecture Risks of the recommended path The architectural principle to take home Key takeaways Right now, somewhere, there's a lakehouse with Highly Confidential in the header. Show me the query pattern. If the pattern is "restrict who opens this," design the workspace. If the pattern is "protect what leaves," design the label. Conflating them is where every thread on that forum starts. And the protection policy sits in between. From the video Demo 1: The Label Changed Nothing Architecture film: Two questions two mechanisms Architecture film: Eight routes onto an item Architecture film: Where protection starts Architecture film: Where label and permission meet Resources Sensitivity labels in Power BI Sensitivity labels troubleshooting Protected sensitivity labels in Fabric and Power BI Roles in workspaces Permission model Row-level security (RLS) with Power BI Information protection in Microsoft Fabric Protection policies in Microsoft Fabric Restrict access to content by using sensitivity labels to apply encryption Export data from a Power BI visualization Apply sensitivity labels to Fabric items Default label policy for Fabric and Power BI Mandatory label policy for Fabric and Power BI Domain-level default sensitivity labels Information protection tenant settings About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    A Label Is Not a Lock

  2. Oct 23

    Five Thousand Files From a Blank Report

    Five Thousand Files From a Blank Report Episode 42 • 2026-10-23 Duration: 10:17 Medallion is the recommended design approach for Fabric. It's also one building block out of five, and the one that sends you a maintenance invoice nobody put on the architecture diagram. What the file-count guardrail means for Direct Lake on OneLake, why schema-enabled lakehouses can't use the automated maintenance activity, and when the whole pattern is overkill. What we discuss How it actually works underneath the abstraction The pattern we keep seeing in the field Where the obvious answer breaks A real Reddit/Microsoft Q&A question unpacked F-SKU realism — what this actually costs When the rejected approach is actually right The concrete recommended architecture The architectural principle to take home Key takeaways Somewhere right now, there's a table sitting at four thousand files on an F64. The price of that answer is maintenance. A schedule, a monitor, and someone who knows where TABLETRAITS lives. Architecture is not religion. Medallion answers one question — what is each stage trusted for — and it leaves the other four to the other four bricks. From the video Architecture film: Five building blocks Architecture film: Each layer is a contract Demo 1: Bronze That Copies Nothing Demo 2: Silver as a Contract Architecture film: When recipe is wrong Resources Understand medallion architecture for Fabric with OneLake OneLake patterns and foundational capabilities Common dbt job patterns Modern data warehouse medallion architecture in Microsoft Fabric What are materialized lake views? Spark SQL reference Optimal refresh Cross-workload table maintenance and optimization What's new Lakehouse Maintenance activity Apache Spark billing and utilization extended capabilities CDF page Data quality in materialized lake views Manage materialized lake views lineage About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    Five Thousand Files From a Blank Report

  3. Oct 22

    Five Doors and One Credential

    Five Doors and One Credential Episode 41 • 2026-09-18 Duration: 8:45 Databricks and Fabric have five integration paths, not one. Each moves data. None moves permissions. We pull apart the one-credential model, the hour-long revoke delay, and the ADLS shortcut that silently bypasses Unity Catalog's row-level security. What we discuss How it actually works underneath the abstraction Where the obvious answer breaks The pattern we keep seeing in the field A real Reddit/Microsoft Q&A question unpacked F-SKU realism — what this actually costs When the rejected approach is actually right Risks of the recommended path The concrete recommended architecture The architectural principle to take home Key takeaways Every path between these platforms moves data. Somewhere right now, a mirrored catalog is serving a workspace that was supposed to see two schemas. And accept that "keep access configurations synchronized" is documentation for there is no sync. Watch the episode The Fabric Friday video on this topic: https://youtu.be/RAqhXwSj_HE From the video 13:33 Demo: Direct Lake Over Databricks Data 3:17 Demo: Mirror a Catalog from Unity Catalog 8:37 Demo: One Credential for Everyone, and a Revoke That Takes an Hour 23:49 Who Pays 17:02 The Reverse Direction Is Three Different Things Resources Mirroring Azure Databricks Unity Catalog Secure Fabric mirrored databases from Azure Databricks Tutorial: Configure Microsoft Fabric mirrored databases from Azure Databricks Enable external data access to Unity Catalog Publish a Unity Catalog catalog to Microsoft Fabric Limitations in Microsoft Fabric mirrored databases from Azure Databricks Enable OneLake catalog federation Integrate OneLake with Azure Databricks Connect Azure Databricks to a OneLake external location Fabric capacity and OneLake consumption OneLake compute and storage consumption Understand your Azure bill for a Fabric capacity Shortcuts in a lakehouse Mirrored Azure Databricks Unity Catalog definition OneLake storage tiers About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    Five Doors and One Credential

  4. Oct 16

    Throttled by Yesterday's Refresh

    Throttled by Yesterday's Refresh Episode 37 • 2026-10-16 Duration: 9:21 Fabric doesn't charge you when work happens — it spreads the cost forward into 2,880 thirty-second timepoints. That one sentence explains every confusing capacity behavior, including why surge protection can reject the SQL it's supposed to protect. What we discuss How it actually works underneath the abstraction Where the obvious answer breaks A real Reddit/Microsoft Q&A question unpacked F-SKU realism — what this actually costs When the rejected approach is actually right Risks of the recommended path The architectural principle to take home Key takeaways And someone's about to halve the SKU because the chart looks calm. The spike they see is a receipt. The smoothed line is the bill that hasn't arrived. Right now, somewhere, there's a capacity chart with a quiet-looking evening. Watch the episode The Fabric Friday video on this topic: https://youtu.be/iE-jdcseCaY From the video 7:42 Demo: Watch One Job Spread 9:51 The Throttling Staircase, and Who Skips Steps 2:50 Demo: Read the Metrics App 12:46 Sizing From the Smoothed Line Resources Fabric capacity throttling policy Smoothing and throttling in Fabric Data Warehouse Throttle triggers and throttle stages Capacity planning guide Autoscale Billing for Spark Fabric operations: interactive and background Understand the metrics app timepoint page Understand the metrics app timepoint summary page Surge protection Capacity and SKUs Burstable capacity in Fabric Data Warehouse Consumption rates and billing for Copilot in Fabric Azure quotas for Fabric capacities About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    Throttled by Yesterday's Refresh

  5. Oct 9

    Your Domain Isn't a Wall

    Your Domain Isn't a Wall Episode 38 • 2026-10-09 Duration: 8:43 Fabric domains look like governance boundaries but the documentation says otherwise. We unpack what a domain actually gives you, the default-domain mechanism that's still running in your tenant, and why the label on the cabinet is not the lock. What we discuss How it actually works underneath the abstraction The pattern we keep seeing in the field A real Reddit/Microsoft Q&A question unpacked Where the obvious answer breaks F-SKU realism — what this actually costs When the rejected approach is actually right Risks of the recommended path The architectural principle to take home Key takeaways There's a filing cabinet in every office with a label on the front. Pattern dictates platform. If the pattern is "organize and find," a domain is the right tool. If the pattern is "restrict and protect," you're in the wrong layer. So here's where I'd land. A domain answers "whose data is this." Use it for discovery and delegation. Build your actual boundaries with workspace roles and item permissions. And audit that default domain list quarterly, because it's still running. Watch the episode The Fabric Friday video on this topic: https://youtu.be/8wRSvnF_CtQ From the video 7:46 Demo: The Contributor Who Cannot Contribute 3:15 Demo: The Filter That Is Not a Fence 12:51 Demo: An Assignment Nobody Made 14:42 The Sensitivity Label Override Ladder Resources Fabric domains Power BI implementation planning: Tenant administration Best practices for domains Domain-level default sensitivity labels Endorsement overview OneLake catalog overview Scope the catalog to a particular domain Governance overview and guidance Roles in workspaces About tenant settings About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    Your Domain Isn't a Wall

  6. Oct 6

    Where Do the Columns Come From?

    Where Do the Columns Come From? Episode 40 • 2026-09-25 Duration: 9:19 Cosmos DB has no schema. Delta tables need one. When mirroring bridges the two, the replicator guesses — and the guess only ever widens. We trace where the bytes come from, why renamed properties leave ghost columns full of nulls, and what stop replication actually does. What we discuss How it actually works underneath the abstraction The pattern we keep seeing in the field A real Reddit/Microsoft Q&A question unpacked Where the obvious answer breaks Risks of the recommended path F-SKU realism — what this actually costs When the rejected approach is actually right The concrete recommended architecture The architectural principle to take home Key takeaways Somewhere right now a Cosmos container is taking writes. And match your Fabric capacity region to the Cosmos account. Three questions before you touch anything. Watch the episode The Fabric Friday video on this topic: https://youtu.be/KJUBmoNFICo From the video 13:19 The Discovered Schema 16:20 Nested Data Arrives as a String 18:47 Stop Is Not Pause 2:53 Demo: Create the Mirror, Count the Rows Resources Mirroring Azure Cosmos DB Cosmos DB mirroring FAQ continuous backup Cosmos DB mirroring limitations Migrate an Azure Cosmos DB account from periodic to continuous backup mode Migrate from Azure Synapse Link to Cosmos DB mirroring Access mirrored data in Lakehouse and notebooks How to: Query nested data Configure private networks for Azure Cosmos DB Fabric Mirroring tutorial Monitor Fabric mirrored database replication Mirroring public REST API Cost of mirroring Mirrored database in Fabric (all sources) Cosmos DB in Microsoft Fabric (native item) About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    Where Do the Columns Come From?

  7. Sep 4

    The Permission Split Inside Read-Only

    The Permission Split Inside Read-Only Episode 36 • 2026-10-02 Duration: 8:03 A workspace carries four boundaries — security, cost, region, lifecycle — and you configured one of them. This episode pulls apart what 'read-only' actually means in Fabric, why OneLake security cannot fence in a Contributor, and the silent security change hiding inside every role promotion. What we discuss How it actually works underneath the abstraction The pattern we keep seeing in the field A real Reddit/Microsoft Q&A question unpacked Where the obvious answer breaks F-SKU realism — what this actually costs When the rejected approach is actually right Risks of the recommended path The architectural principle to take home Key takeaways Somewhere right now, someone's promoting a colleague to Contributor so they can fix a report. Show me the query pattern. That tells you whether Viewer plus OneLake security is enough, or whether you need to split. Here's what I'd carry out of this. Every workspace is a cut through the lake, and cuts carry consequences. Security, cost, region, lifecycle — you're deciding all four whether you chose them or not. Watch the episode The Fabric Friday video on this topic: https://youtu.be/tpePwivxtzk From the video 6:25 Demo: Same Table, Two Routes 12:28 Demo: Watch the Override 14:35 Who Can Claim Your Workspace Resources Roles in workspaces Permission model Give users access to workspaces OneLake security access control model Capacity reassignment restrictions Understand Fabric licenses and capacity Manage workspaces What is a lakehouse Reassign a workspace to a different capacity Create a workspace Workspace roles in Fabric Data Warehouse Workspace roles in Lakehouse Roles in workspaces in Power BI Build permission for shared semantic models Get started with Microsoft Fabric About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    The Permission Split Inside Read-Only

  8. Aug 28

    The Snapshot Locked in a Room

    The Snapshot Locked in a Room Episode 35 • 2026-08-28 Duration: 8:42 Graph in Fabric materializes a second copy of your data for traversal. That copy explains the staleness, the rebuild cycle, and the workspace boundary that decides whether graph belongs in your architecture. What we discuss How it actually works underneath the abstraction The pattern we keep seeing in the field A real Reddit/Microsoft Q&A question unpacked Where the obvious answer breaks The concrete recommended architecture F-SKU realism — what this actually costs When the rejected approach is actually right Risks of the recommended path The architectural principle to take home Key takeaways And somewhere right now, there's a queryset returning a result that was true yesterday. A query timeout of twenty minutes. That's a ceiling you can plan around. What you can't plan around is the copy. It's stale between rebuilds, locked to the workspace, and if you query it carelessly it hands you a partial answer with a warning you didn't read. Accept those three things and it solves relationship... From the video Architecture film: Snapshot not view Architecture film: Four ways to ask Architecture film: Staleness chain Demo 3: Change the Model, Watch the Queryset Architecture film: The ceilings Resources How graph in Microsoft Fabric works Current limitations of graph in Microsoft Fabric Troubleshooting and FAQ for graph Quickstart: Create your first graph Tutorial: Introduction to graph Tutorial: Query the graph by using the query builder Example graph datasets Optimize GQL query performance GQL language guide GQL quick reference GQL query API (REST) Graph data models Design a graph schema Fabric Data Agent OneLake security About the show AI-generated voices. Matthias — cloned voice. Fabia — designed AI co-host. See Matthias live on YouTube (Fabric Friday), at his meetups, and at conferences like FabCon. Hosted by Matthias Falland — Microsoft Data Platform MVP and community architect behind the Fabric Periodic Table. New episodes every Friday. Submit your case Have an architecture decision you are wrestling with? DM Matthias on LinkedIn — find him as Matthias Falland. Three to five sentences about the decision, your team size, and your current stack. We anonymize before airing. This podcast was generated by AI. Both voices are synthetic: Matthias is a cloned voice, Fabia is a designed AI co-host. Brand design based on fabricperiodictable.com.

    The Snapshot Locked in a Room
See All (43)

Trailer

About

Architecture decisions for Microsoft Fabric. Anonymized real customer scenarios, cost realism, counter-arguments included. Weekly episodes aligned with Fabric Friday recordings.

Information

You Might Also Like