A pipeline defines what to read, how to map it and where imported records belong. Use the pipeline editor’s validation and release controls so configuration is reviewable before it processes shared data.
Publishing or running a pipeline can create or update shared records. Review the destination and source scope before confirming.
Before you start
You need administration access, a working source connection and a destination object. Know which source fields identify records consistently.
Use Settings search or the navigation to find the controls you need.
Configure and validate the import
Open Settings → Data sync and create or open the intended pipeline.
Select the connected source and the supported read operation. Review parameters, filters and the expected record shape.
Map source identifiers and fields into the destination object. Check compatible field types, units and relationships.
Use the editor’s preview and validation steps. Review representative rows and resolve all required mapping issues.
Review the destination, scope and schedule with the owner of the source data before publishing or running the pipeline.
Use the product’s publish and run controls, then inspect the run result and the actual destination records.
Check the result
The intended version is published, the run completes and representative records have the expected identifiers, values and relationships. Saving a draft or passing a preview alone does not establish a successful sync.
Validate meaning as well as shape
A date field can be technically compatible but represent the wrong event. Check requested versus confirmed dates, open versus total quantities, currency and units with the business owner.
For related objects, confirm the referenced records exist and the relationship uses the right key. Test a small, representative scope before broadening coverage.
When a provider changes
If an operation or field disappears, reopen the pipeline and use supported replacements. Resolve the editor’s diagnostics before publishing. Use the visible product validation path for changes to production configuration.
If something does not look right
What you see | What to check |
Preview returns no rows | Check connection, source filters, operation parameters and available source data. |
Validation blocks publication | Resolve each required mapping or schema issue, then validate again before running. |
Is a successful preview a completed import?
Is a successful preview a completed import?
No. Verify the published version, the actual run and its destination records separately.
Continue your work
Open the relevant area in Cartograph to continue with your workspace data.