Based on the available knowledge sources, there is currently no explicit configuration option in Elementary to disable showing failed records (the sample of failing rows) in the web UI. There are discussions about controlling the sample size of failed records using the test_sample_row_count variable, which determines how many failed rows are stored and displayed (default is 5), and there have been user requests for the ability to generate reports without including sample data for privacy or performance reasons. However, as of now, there is no documented config to completely hide or disable the display of failed records in the UI or reports. The feature to skip fetching sample data is not yet available, though it has been discussed as a potential future enhancement [Slack discussion on skipping sample data](https://slack.com/archives/C02CTC89LAX/p1669054860106729). If you want to limit the number of failed records shown, you can set test_sample_row_count to a lower value (even 0), but this will affect both storage and display. There is no config to selectively hide them only in the UI while still storing them elsewhere [Sample size config discussion](https://slack.com/archives/C02CTC89LAX/p1704319265407929).

The ability to configure the number of failed test rows stored and displayed using the test_sample_row_count variable has been present in Elementary for several versions. According to the knowledge sources, this configuration was already available in Elementary version 0.7.3, as referenced in a Slack discussion from February 2023, where users were advised to upgrade to at least 0.7.3 to benefit from the separate test_result_rows table and the sample size configuration [Slack: test_sample_row_count and version 0.7.3](https://slack.com/archives/C02CTC89LAX/p1678288875135819). There is no specific information in the knowledge sources about the minimum required dbt-core or dbt-databricks versions for this feature. The documentation and community discussions focus on the Elementary package version and do not mention any dbt-core or dbt-databricks version dependencies for the test_sample_row_count configuration. If you need to ensure compatibility, it is recommended to use at least Elementary version 0.7.3 or later. For dbt-core and dbt-databricks, please refer to the official compatibility matrix or documentation, as the knowledge sources do not provide explicit details on this aspect.

Yes, if you set test_sample_row_count: 0 in your dbt_project.yml, Elementary should not include any sample rows of failed records in the test_result_rows table or in the result_rows column of elementary_test_results. This effectively disables the storage and display of failed row samples in reports and the UI. This configuration is specifically mentioned as the way to opt out of collecting and displaying failed row samples for privacy or other reasons [Security and Privacy Docs](https://docs.elementary-data.com/cloud/general/security-and-privacy#what-information-does-elementary-collect%3F). However, please note that in some older versions, there were reports that setting test_sample_row_count: 0 did not fully disable the sampling mechanism, and a fix was planned for this behavior. If you are using a recent version (0.7.3 or later), this configuration should work as expected [Slack: test_sample_row_count and version 0.7.3](https://slack.com/archives/C02CTC89LAX/p1678288875135819) [Slack: Disabling samples](https://slack.com/archives/C02CTC89LAX/p1672963194187039). If you still see sample rows after setting this variable, make sure you are using a supported Elementary version and that the variable is placed directly under vars: in your dbt_project.yml, not under models or other sections [Slack: Correct placement of test_sample_row_count](https://slack.com/archives/C02CTC89LAX/p1704277382272089).

Elementary does not provide a built-in command equivalent to dbt docs serve for serving the edr report as a local web server. When you run edr report, it generates a static HTML file (by default at edr_target/elementary_report.html), which you can open directly in your browser. If you want to serve the report locally over HTTP, you would need to use a separate tool, such as Python’s built-in HTTP server, to serve the directory containing the report file. There have been user requests for a built-in serve feature, but as of now, this is not available. The recommended approach is to open the HTML file directly or use your own method to serve it locally [Slack: serving the report](https://slack.com/archives/C02CTC89LAX/p1669034786609089) [Report generation and viewing](https://docs.elementary-data.com/oss/guides/generate-report-ui).