> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cognite.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Monitor extraction pipelines

> Monitor extraction pipelines to ensure reliable data integration and quickly identify and troubleshoot failures.

Improve data quality management and make sure the data integration into Cognite Data Fusion (CDF) [data sets](/cdf/data_governance/concepts/datasets/index) 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

<Steps>
  <Step title="Navigate to extraction pipelines">
    Navigate to <span class="ui-element">Data fusion</span>.

    * To view all extraction pipelines across data sets in a CDF project and their latest status, navigate to <span class="ui-element">Integrate</span> > <span class="ui-element">Extraction pipelines</span>.
    * To view extraction pipelines for the **selected data set**, navigate to <span class="ui-element">Data catalog</span>. Then select a data set, and open the **Lineage** tab.
  </Step>

  <Step title="Open pipeline details">
    Select the pipeline name in the list to open the **Extraction pipeline overview**. Depending on the extractor, open the **Run history** tab (typical for on-premises extractors) or the **Insight** tab (typical for hosted extractors) to view run-level details.
  </Step>
</Steps>

After you open a pipeline's details, use the following to interpret pipeline status, uptime, per-run messages, and where to investigate next.

## Understanding status and uptime

The **Extraction pipeline** statuses reflect the latest known state of the pipeline and its runs. Treat the status as a signal to investigate. One failing step, such as authentication, connectivity, scheduling, or a failed upload to CDF, can show **Failure** or otherwise look unhealthy even when the rest of your integration is still healthy.

Choose the tab that matches your extractor.

<Tabs>
  <Tab title="On-premises extractors">
    On **Overview**, **Last reported status** summarizes successful and failed critical operations. **Last connected** shows when an extractor last communicated with CDF, and **Schedule** shows when the extractor is expected to run.

    On **Run history**, the **Message** column shows the errors or success summaries. Compare timestamps and **Configuration revision** across runs. Treat any failing status or stale timestamp as a signal to investigate.
  </Tab>

  <Tab title="Hosted extractors">
    On **Overview**, each row in the jobs table is a **job** with an MQTT **topic filter** (the broker subscription in the **Topic filter** column). Each job has its own **Status** (for example **Running** or **Paused**). **Topic filters status** aggregates how reliably **all topic filters on this extractor** receive traffic over a recent window (uptime percentage and chart; broker subscription and ingestion health, not full validation in CDF).

    **Insight** or **View full insight** shows a deeper view of uptime and run-level signals.
  </Tab>
</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**.

Choose the tab that matches your extractor.

<Tabs>
  <Tab title="On-premises extractors">
    * Select the pipeline name to open the **Extraction pipeline overview** tab to find documentation and key properties.
    * Use **Owner** and **Contacts** to see who to inform about a failure.
    * Review and confirm what is happening with **Last reported status**, **Last run time**, and **Last connected**.
    * Switch to the **Run history** tab to review past runs, filter by time and **Last run status**.
    * Review the chart, extractor metrics, and logs to get broader knowledge to see how the extraction process has performed over time.
    * When you have fixed the failure, update the pipeline documentation with any lessons learned.
  </Tab>

  <Tab title="Hosted extractors">
    * Review extractor details, connection details, and recent errors.
    * Compare each job's **Status** with **Topic filters status** and review the uptime to measure availability in operations.
    * Use **Insight** (or **View full insight**) for charts and detail, and review extractor metrics or logs if data in CDF still looks wrong.
  </Tab>
</Tabs>