Pipeline Console

The Pipeline Console is a pre-built SAPUI5 freestyle app (flexible column layout) for inspecting and operating pipelines at runtime. It reads and writes the management OData API at /pipeline/ — list pipelines, inspect source and target data, review run history, trigger runs, and manage in-process schedules.

The console ships inside the cds-data-pipeline npm package at app/pipeline-console/ (pre-built static assets). The UI runtime uses SAPUI5 (including sap.suite.ui.commons for the data-flow network graph); the self-contained build is subject to the SAP Developer License, not the Apache 2.0 license of OpenUI5.

What you get

Surface	Path	Notes
Pipeline Console (this page)	/pipeline-console/	Freestyle FCL — list / detail / data inspector. Shipped with the plugin.
Management OData API	/pipeline/	Opt in via management.reuse.api or manual CDS import. See Management service.

The console manifest binds its default OData model to /pipeline/ (absolute URL). UI and API share the same CAP origin — no CORS or reverse-proxy path rewriting when both run on the same server.

Layout

Column	Content
Begin — Data Pipelines	Pipelines tab: searchable list with status summary, last-run column, error counts, and manual refresh. Overview tab: landscape network graph grouped by service (entity / consumption-view nodes inside each group)
Mid — Pipeline detail	Overview (status, health counters, cumulative statistics, raw JSON); Runs; Configuration (coded / override / effective); Data flow graph; Data Preview. Orange marks custom hooks or non-default configuration
End — Data inspector	Source/target toggle, column selection, advanced filters, and paged preview table via inspectData

Screenshots

Landscape — pipelines grouped by remote service; entity nodes and flow arrows show source → pipeline → target:

Pipelines list and run history — schedules, status, and per-run statistics (including federation-bound replicate jobs):

Status semantics

Pipeline status (idle / running / failed) reflects the tracker lifecycle — whether a run is in progress and whether the last run failed — not data freshness. Last successful run (lastSync) is written only after a completed successful run; until then the UI shows Never run yet. Enabled / Paused gates scheduled ticks only — manual Run pipeline still works while paused.

The console polls every 30 seconds when idle and every 3 seconds while any pipeline is running, refreshing the pipeline list and the Overview and detail data-flow graphs.

Console capabilities

Actions call the same bound and unbound operations documented in Management service: start, setSchedule, clearSchedule, setOverrides, clearOverrides, setEnabled, configView, flush, inspectData, flowMetadata, and landscapeMetadata. Configuration layering is described in Configuration overrides.

Quick start — local reuse (recommended)

Add to your project's package.json:

{
  "cds": {
    "requires": {
      "data-pipeline": {
        "impl": "cds-data-pipeline",
        "management": {
          "reuse": {
            "api": true,
            "console": true
          }
        }
      }
    }
  }
}

Then:

cds deploy
cds watch

Open http://localhost:4004/pipeline-console/index.html (port from your profile).

No server.js changes, no local UI copy. See Feature activation for reuse vs own decision rules.

Developing the UI (TypeScript)

Edit sources under packages/cds-data-pipeline/app/pipeline-console-src/webapp/ only. The dev server transpiles TypeScript on save and proxies requests under /pipeline to the contributor dev backend on port 4100.

# Terminal 1 — dev backend (four pipelines)
bash packages/cds-data-pipeline/examples/_dev/pipeline-console/start.sh
# or: cd packages/cds-data-pipeline && npm run dev:console-backend

# Terminal 2 — live TypeScript UI
cd packages/cds-data-pipeline && npm run dev:pipeline-console

Open http://localhost:8090/index.html. Pipeline details: examples/_dev/pipeline-console/README.md.

Script	Purpose
dev:console-backend	Start dev backend on :4100
dev:pipeline-console	Live TS UI on :8090 (proxies to :4100)
build:pipeline-console	Production bundle into app/pipeline-console/

Feature examples (01–07) use the built console at /pipeline-console/ — not this dev flow.

BTP production — project-owned UI

For HTML5 Application Repository + approuter deployments:

"data-pipeline": {
  "impl": "cds-data-pipeline",
  "management": { "reuse": { "api": true } }
}

cds add pipeline-console
cds add html5-repo
cds add mta
cd app/pipeline-console && npm install && npm run build:cf

Use --source on cds add pipeline-console to copy TypeScript sources from app/pipeline-console-src/ instead of the pre-built bundle when you need to customize the UI.

Do not set management.reuse.console when you already ran cds add pipeline-console.

Manual management API import

When you prefer explicit CDS control over management.reuse.api:

using from 'cds-data-pipeline/index.cds';

Mount the console with cds add pipeline-console (BTP) or management.reuse.console (local reuse).

Composite / mashup apps

Pipeline services are embedded when your CDS model reaches the plugin definitions. With management.reuse.console, the plugin mounts static UI assets at /pipeline-console/ during bootstrap. Dynamic requests still hit /pipeline/ on the same process.

Security

The plugin does not attach @(requires: …) to the management service. Secure /pipeline in your CDS model and deployment (XSUAA, approuter, annotations). The console uses the same OData session as the rest of your app. inspectData exposes live source and target rows — treat it with the same authorization as the underlying entities.

Production hardening built into the pipeline inspector:

Control	Behavior
Opt-out	Set management.inspect: false on your cds-data-pipeline requires entry to disable inspectData and hide inspector tabs (inspectCapabilities returns none for both sides).
Field / entity exclusion	Honors @HideFromDataInspector on modeled source/target entities — hidden elements are never selected; entity-hidden sides report none and return no rows.
Audit logging	When @cap-js/audit-logging is installed (cds.requires['audit-log']), preview reads emit SensitiveDataRead for columns annotated with @PersonalData.IsPotentiallySensitive (best-effort; failures are logged, never block the preview).

For app-wide entity browsing, search, and Work Zone tiles with full compliance wiring, use @cap-js/data-inspector in your project. The pipeline inspector remains the right tool for pipeline-scoped previews (remote OData/REST sources, secondary DB targets, custom adapters) that the cap-js plugin cannot see.

See Securing /pipeline.

Without a UI

The management API is fully usable without any UI:

GET /pipeline/Pipelines
GET /pipeline/PipelineRuns
POST /pipeline/execute
GET /pipeline/Pipelines('MyPipeline')/DataPipelineManagementService.inspectData(side='target',top=50,skip=0)

See Management service for the full surface.

See also

• Feature activation — decision tree, reuse vs own, duplicate-loading rules
• Get started — end-to-end walkthrough
• Programmatic API — inspectData, getFlowMetadata, and the methods the console calls
• Management service — OData entities, actions, hooks
• Feature catalog — observability and UI capabilities