Managing Workspaces
Create completely isolated organizational Workspaces equivalent conceptually to unique 'Projects'. Restrict and keep experimental staging pipelines entirely separated from mission-critical production runtime architectures.
Environment Isolation
Dynamically building and testing a newly architected, highly destructive pipeline transformation against a live production transactional database is universally considered a terrible, fireable error.
Hard Isolation Architecture
Workspaces logically solve this parameter conflict. A Workspace mathematically behaves as an entirely disparate tenant instance, seamlessly replicating your DataFlow AI design topology while securely enforcing strict Environment Variable distincts (e.g. mapping strictly to db_secret_dev vs db_secret_prod).
Cross-Workspace Promotion Integrations
A standard high-fidelity CI/CD software development framework maps beautifully onto Workspaces.
Development Workspace
Engineers organically connect to localized sample databases and heavily mocked API endpoints natively. The intelligent AI Copilot securely generates the initial raw DAG geometry safely here, allowing endless iteration without structural risk constraints.
Staging Workspace
JSON Configurations are merged effortlessly via underlying Git branch mapping. DataFlow AI mathematically points the data connectors dynamically towards the secure Staging Snowflake environment. Complex automated testing frameworks and remote Great Expectations QA validation scripts securely execute mathematically here.
Production Workspace
Once successfully verified across all tests, the unified declarative code branches seamlessly deploy towards Prod. This highly restricted state is securely locked to verified Administrator or specific Engineering approval layers only, deploying the logical graph processing directly against multi-billion row live server architectures.
© 2026 DataFlow AI Docs