Based on the knowledge sources, there are several approaches to improve the performance of Elementary Data in your dbt test runs: ## Performance Optimization Options 1. Disable Column Autoupload: Setting

disable_dbt_columns_autoupload: true

in your variables can significantly reduce processing time. One user reported that with column uploading enabled, their process was taking almost 30 minutes, and disabling it resolved the issue [Slack forum discussion](https://slack.com/archives/C02CTC89LAX/p1707997271403599). 2. Disable Artifacts Autoupload: You can set

disable_dbt_artifacts_autoupload: true

in your project variables to reduce the processing overhead [Slack forum discussion](https://slack.com/archives/C02CTC89LAX/p1705567419922439). 3. Adjust Test Sample Row Count: Setting

test_sample_row_count: 0

can improve performance. This will mean you won't see sample rows for failed tests, but it can reduce runtime [Slack forum discussion](https://slack.com/archives/C02CTC89LAX/p1695649576308329). 4. Use Chunked Inserts: For large datasets, you can instruct Elementary to upload results in chunks by adding these variables to your `dbt_project.yml`:

vars:      insert_rows_method: chunk      dbt_artifacts_chunk_size: 500

[Slack forum discussion](https://slack.com/archives/C02CTC89LAX/p1703928287718209) 5. Disable Failed Count Calculation: Setting

calculate_failed_count: false

can speed up processing, though you'll lose the count of failed results [Slack forum discussion](https://slack.com/archives/C02CTC89LAX/p1695328060256599). ## Multi-threading Elementary does utilize the threading configuration from your dbt profile. The tests run in parallel in a threaded manner, though some of the logging might make it appear sequential [Slack forum discussion](https://slack.com/archives/C02CTC89LAX/p1658915396893639). You might want to check your

profiles.yml

to ensure you have an appropriate thread count set (e.g.,

threads: 8

). ## Additional Considerations - The on-run-end hooks are a known performance bottleneck, especially with larger projects - Upgrading to newer versions of Elementary might help, as performance improvements are regularly released - Consider running Elementary artifacts upload separately if you only need them for specific purposes:

dbt run --select elementary.edr.dbt_artifacts

[Slack forum discussion](https://slack.com/archives/C02CTC89LAX/p1717515678662739) If these optimizations don't provide sufficient improvement, you might want to consider opening an issue on the Elementary GitHub repository with details about your specific setup for more targeted assistance.

I don't see this referenced in the docs @Noya Offer, is this the same as

disable_dbt_columns_autoupload: true

? https://docs.elementary-data.com/dbt/on-run-end_hooks#disable-metadata-models-updates