NOTES
Useful things,
without the content treadmill.
Architecture, migrations, platform engineering, cloud cost and AI — written when there is something concrete worth sharing.
Building a metadata-driven Databricks workspace migration accelerator
The design controls that turn notebook/job copying into an actual migration capability: inventory, dry runs, destination validation, dependency checks, auditability and rollback.
Synapse vs Databricks: the decision is bigger than compute
A practical comparison framework for architects who need to make a business case, not just compare benchmark numbers.
Where Databricks cost quietly leaks
The cost patterns that often hide behind healthy-looking cluster and warehouse dashboards.