Skip to main content

20.6 APIs, CLI & CI/CD

SAP Datasphere supports programmatic administration and modeling scenarios through APIs and the Datasphere command-line interface for supported operations. Data Flow documentation explicitly documents command-line management of modeling objects.

CI/CD architecture​

Git or change source -> CI runner -> Datasphere API/CLI -> TEST -> validation -> approval -> PROD

Pipeline stages​

Stage 1 — Validate​

Check:

  • syntax
  • naming
  • object references
  • expected artifacts
  • package completeness

Stage 2 — Deploy to TEST​

Use an isolated service identity with only required privileges.

Stage 3 — Validate​

Run metadata, security and data tests.

Stage 4 — Promote​

Require approval for production.

Stage 5 — Observe​

Capture deployment result, runtime health and reconciliation.

Secrets​

Never place credentials in source control or notebooks. Use the CI platform's protected secret mechanism and rotate credentials according to enterprise policy.

Idempotency​

A deployment should be safe to repeat where the underlying operation supports it. Pipelines must distinguish:

  • create
  • update
  • replace
  • destructive change

Architect decision​

Prefer APIs for repeatability, but keep a documented manual recovery path. Automation that cannot be debugged by operators is an operational liability.