Skip to main content
A destination provides the CDF credentials and optional data set scope that hosted extractor jobs use to write data. What gets written (time series, events, RAW table rows, data model instances, or Records in streams) is determined by the job’s format or, for custom format jobs, its mapping. See Message formats for built-in format options. You configure destinations to control how extracted data is stored, organized, and governed within CDF.

​
What destinations are

A destination contains a CDF session (credentials) and an optional targetDataSetId that scopes all data written through it to a specific data set. Multiple jobs can share one destination; when they do, their requests are combined, reducing the number of calls made to CDF APIs while metrics are still reported per job.

​
Creating and managing destinations

You create a destination by providing an externalId and CDF credentials. Once created, you reference the destination by its externalId when creating a job. You can update the credentials or data set ID, and delete a destination when it is no longer associated with any jobs.
A destination session stores the capabilities of the user or client credentials at creation. Changing those capabilities later does not update the destination. Updating credentials on the destination is not enough. Create a new destination, point jobs at it, then delete the old destination.
Grant destination credentials access only to the CDF resources the extractor needs to write. See Hosted extractors overview for the required capabilities per output type.
Last modified on August 28, 2026