Platform

Solutions

Resources

Pricing

Try

February 26, 2026

Full Court Press release ๐Ÿ€๐Ÿ”ฅ

Bring the heat. Push the pace. Leave it all on the floor.

The Full Court Press release strengthens performance, reliability, and developer workflow efficiency in Mage Pro. This update is about keeping your pipelines, environments, and deployments ready to respond when workloads pick up pace.

Like a team running a relentless full court press, your infrastructure should stay active, resilient, and hard to break under pressure.


๐Ÿงฌ Reproducible, isolated dependency environments

Dependency management in Mage Pro has been redesigned to keep runtime environments clean and predictable.

Instead of installing packages globally, dependencies now live inside isolated virtual environments that are created when workspaces start.

Each environment is fingerprinted using Python version, requirements, and tooling state. Once built, environments are immutable and safely reusable across pods. Builds are executed with concurrency-safe locking, real-time install streaming, and automatic fallback support when switching between uv and pip workflows.

Teams can also manage environments using CLI tools like mage deps ensure, mage deps install, and mage deps list to inspect fingerprints, timestamps, and runtime metadata. Cleanup and retention control are available through mage deps cleanup.

Environment state is stored in .mage_env.json, making dependency behavior reproducible across clusters.

๐Ÿ“– Documentation: https://docs.mage.ai/development/dependencies/dependency-management#dependency-management

โ˜ธ๏ธ Kubernetes defaults that scale

Workspace execution on Kubernetes can now inherit configuration defaults from simple config files.

Teams can standardize executor settings, resource limits, and placement rules like tolerations or node selectors without repeating configuration across environments. New workspaces automatically inherit these defaults, helping reduce configuration drift from development to production.

This improvement makes multi-tenant cluster behavior more predictable and easier to operate.

๐Ÿ“– Documentation:

https://docs.mage.ai/guides/developer-ux/workspaces#default-workspace-config-file

https://docs.mage.ai/production/executors/k8#default-config-file

โšก BigQuery Storage Write API: Faster & Safer Streaming

Streaming pipelines targeting Google BigQuery now use the Storage Write API for higher throughput and safer execution.

Arrow-based schema handling improves type consistency, subprocess execution prevents gRPC deadlocks, and cancellation logic is handled at the process group level. If streaming runs fail, new runs are automatically created to keep data flowing.

๐Ÿ“– Documentation: https://docs.mage.ai/guides/streaming/destinations/bigquery

๐Ÿข First-Class Microsoft Fabric Support

Microsoft Fabric is now fully supported inside Mage Pro through integrations from Microsoft.

Developers can write and run Fabric SQL inside pipeline blocks, use Fabric as a data source, and export results as part of orchestration workflows. The SQL editor understands TransactSQL syntax, providing better highlighting and autocomplete.

Fabric pipelines are now connected from configuration through execution.

๐Ÿ“– Documentation: https://docs.mage.ai/integrations/databases/MicrosoftFabricWarehouse

๐Ÿง‘โ€๐Ÿ’ป Workspace & Deployment Reliability

We strengthened core operational surfaces across workspaces and deployments:

  • More robust Kubernetes workload creation behavior

  • Improved workspace config handling (including empty container configs)

  • Detailed Kubernetes job error logging

  • Deployment app stability improvements (including handling folder copy edge cases)

These changes reduce deployment friction and make cluster-level behavior more transparent.

๐Ÿ”„ Data Integration & CDC Stability

This release improves reliability across batch and streaming integrations:

  • Boolean column sync fixes in Postgres CDC

  • Delta Lake compatibility updates with stricter config validation

  • Safer error logging to prevent credential exposure

  • Great Expectations block fixes

  • Improved UI state handling for data integration blocks

Data pipelines are now more secure, predictable, and compatible across version upgrades.

๐ŸŽจ UI & Developer Experience Refinements

We continued refining the development experience with:

  • Light mode fixes across editor and settings pages

  • File browser search and layout improvements

  • Format JSON button and improved coder panel behavior

  • More transparent stack traces with additional internal paths

  • Improved error messages for invalid catalog errors

  • Optional hidden healthcheck logs

  • Default schedule improvements

  • Scheduler memory safeguards

The result is a cleaner interface and better debugging clarity during daily development.

๐Ÿ”ฅ Feel the pressure. Play faster.

The Full Court Press release keeps environments reproducible, streaming pipelines efficient, and deployments ready for the next possession.

Update your Mage AI Pro workspace and keep the defense on its heels. ๐Ÿ€

Prologs