Based on the knowledge sources, I can explain the difference between

disable_dbt_artifacts_autoupload

and the

edr

CLI approach: ## Key Differences

disable_dbt_artifacts_autoupload

variable: - This variable disables the automatic upload of dbt artifacts (project metadata like

dbt_models

,

dbt_tests

,

dbt_sources

, etc.) during the

on-run-end

hook - When set to

true

, the hook will still upload run results and test results from each dbt invocation - You need to manually run

dbt run --select elementary.edr.dbt_artifacts

to update the artifacts when your project changes

edr

CLI approach: - When you run

edr

(for monitoring or reporting) from your dbt project directory, it automatically uploads the dbt artifacts before generating reports or sending alerts - This creates an overlap with the hook's artifact upload functionality ## Recommended Setup According to [a discussion about concurrent runs](https://slack.com/archives/C02CTC89LAX/p1672966052599339), the recommended approach is: 1. Set

disable_dbt_artifacts_autoupload: true

in your

dbt_project.yml

2. Let

edr

handle the artifact uploads when you run monitoring/reporting 3. The

on-run-end

hook continues to upload run results and test results from each invocation This setup reduces load on your dbt invocations while keeping the hooks lean, as explained in [this thread about reducing hook overhead](https://slack.com/archives/C02CTC89LAX/p1672966052599339). The artifacts (project metadata) don't need to be uploaded on every invocation since they only change when your project structure changes, while run/test results are specific to each invocation.

Based on the knowledge sources, I can clarify the difference between disabling Elementary models with

edr: +enabled

and using the disable variables for hooks: ## Key Differences Disabling Elementary models (

edr: +enabled: false

): - Prevents Elementary models from being created/updated during

dbt run

- According to [this Slack discussion](https://slack.com/archives/C02CTC89LAX/p1673628649610999), when you disable the package using

enabled: false

, the hooks are disabled and models are not created - This means test results will NOT be collected because the hooks won't function properly without the underlying models Using disable variables (recommended approach): - Variables like

disable_run_results

,

disable_tests_results

,

disable_dbt_artifacts_autoupload

, and

disable_dbt_invocation_autoupload

control what the hooks upload - According to [the same discussion](https://slack.com/archives/C02CTC89LAX/p1673628649610999), using these vars "will disable the

on-run-end

hook upload. The package models will still be part of your

dbt run

, and will be empty as there is no hook running to load data into them." ## Your Use Case If you want to disable model updates but keep test results collection, the knowledge sources indicate this creates a problem. As explained in [this thread](https://slack.com/archives/C02CTC89LAX/p1669625539486519), when Elementary models are disabled with

+enabled: false

, it causes issues with test results collection. The [recommended approach from the FAQ](https://docs.elementary-data.com/oss/general/faq) is to: 1. Keep Elementary models enabled but exclude them from regular runs using

--exclude elementary

2. Run Elementary models explicitly only when needed (after installation or upgrades) Alternatively, you can use the var-based approach:

yaml models:   elementary:     +schema: elementary     edr:       +enabled: "{{ var('enable_elementary_models', false) }}"

Then run:

dbt run --select elementary --vars '{"enable_elementary_models": true}'

only when upgrading. However, the hooks will still work to collect test results even when models are disabled this way, as long as the models exist in your warehouse from a previous run.

Based on the knowledge sources, I need to clarify an important point about disabling the

edr

module path: ## The Issue with Disabling Elementary Models According to [this Slack discussion](https://slack.com/archives/C02CTC89LAX/p1680689840547609), when Elementary is set to

enabled: false

, "it's like it is no longer installed in your project. Nothing will work on runs when it's disabled - the results logging, the tests, the metadata collection." Similarly, [another discussion](https://slack.com/archives/C02CTC89LAX/p1673628649610999) confirms that when you disable the package using

enabled: false

, "the hooks are disabled and models are not created." ## The Answer to Your Question No, if you disable the

edr

module path with

+enabled: false

, the run results and test results will NOT be uploaded. The

on-run-end

hooks depend on the Elementary package being enabled to function properly. ## The Recommended Solution According to [the FAQ](https://docs.elementary-data.com/oss/general/faq), if you want to prevent Elementary models from running during regular

dbt run

commands while still collecting results, you should: 1. Keep the models enabled but exclude them from runs using

--exclude elementary

2. Use the disable variables to control hook behavior if needed:

yaml vars:   disable_run_results: "{{ target.name != 'prod' }}"   disable_tests_results: "{{ target.name != 'prod' }}"   disable_dbt_artifacts_autoupload: "{{ target.name != 'prod' }}"   disable_dbt_invocation_autoupload: "{{ target.name != 'prod' }}"

This way, the hooks will continue to work and collect results, but you control when the Elementary models themselves are updated.