Thoughts on data engineering, cloud architecture, and technology. 2171 articles and counting.
Showing 24 of 2171 posts — page 1 of 91
Item-level CI/CD in Microsoft Fabric is a settled capability. The unsolved problem is source of truth and tenant governance as code — the workspaces, capacity, domains, tenant settings, connections, and roles that Git integration does not manage. A blueprint for tenant-as-code.
The Fabric admin portal is one interface over the governance model, not the model itself. How to build governance-as-automation on the same admin, scanner (WorkspaceInfo), and REST APIs a service principal can call directly — with least-privilege identity separation as a first-class design constraint.
A shortcut splices two independently governed permission domains together. Treating it as an access-control and identity problem: two-layer permissions, pass-through vs delegated auth, the Direct Lake / SQL identity-passthrough exception, and where enforcement silently changes hands.
Synapse serverless SQL infers long text columns as varchar(8000), and you can't fix it from PySpark. Here's why it happens — and how to fix it properly across Spark, SQL, and Power BI.