Salt la conținutul principal

Monitor extraction pipelines

Improve data quality management and make sure the data integration into Cognite Data Fusion (CDF) data sets is reliable by monitoring extraction pipelines. If any data integration issues occur, you can be quickly notified, start troubleshooting, and contact the pipeline owner and other stakeholders.

View extraction pipeline details

  1. Navigate to Data management.

    • To view all extraction pipelines across data sets in a CDF project and their latest status, navigate to Integrate > Extraction pipelines.
    • To view extraction pipelines for the selected data set, navigate to Data catalog. Then select a data set, and open the Lineage tab.
  2. Select the pipeline name in the list to open the Extraction pipeline overview and Extraction pipeline run history tabs.

Troubleshooting failed integrations

If an extraction pipeline fails or a continuous data flow into CDF is interrupted, you'll get an email from the system (if activated in setup). Select the email link to open the Extraction pipeline overview. On this page, the Last status column indicates successful and failed critical operations. Select the Failure flag to see the error message from the extractor. The Last connected column shows when an extractor last communicated with CDF, and the Schedule field shows when the extractor is expected to run.

Select the pipeline name to open the Extraction pipeline overview tab to find all documentation and key properties for a pipeline. The documentation added should give you context and insights useful for troubleshooting the pipeline, and the names in the Owner and Contacts fields let you know who to inform about the failure.

Switch to the Extraction pipeline run history tab to find historical runs that will help you understand past and present behavior of the pipeline. You can filter on time, run status and search messages to get broader knowledge about how the extraction process have performed over time.

When you have located and corrected the failure, make sure to update the documentation with any lessons learned.