The cloud-native orchestrator for the whole development lifecycle, with integrated lineage and observability.

dagster

Hi...Does the `FreshnessPolicy` &amp; `build_asset_reconciliation_sensor`  as outlined apply to dbt Cloud assets? I'm aware that if you select individual assets to materialize, Dagster will override the dbt cloud job with that selection, so I'm wondering Dagster takes the same approach with freshness and reconciliation.

E.g. a non-dbt cloud asset, A, is stale and dbt cloud asset, B, is dependent on A.  Asset B has dbt children C &amp; D but dbt cloud job also has E, F and G which has no dependencies). Will the reconciliation sensor and FreshnessPolicy work and result in Dagster triggering a dbt build for only B, C and D?
<https://docs.dagster.io/guides/dagster/scheduling-assets>

Hi <@U01JXQCBR29>! because dbt Cloud assets now support subsetting, the sensor should work as you're expecting (i.e. only kicking off runs of the assets required to meet the freshness policy) :slightly_smiling_face:

<@U01J51Y6B9D> Am I correct to assume that although the sensor would work via other criteria, I can't actually apply a FreshnessPolicy to dbt Cloud assets as there's doesn't seem to be support for this in `load_assets_from_dbt_cloud_job`

Example:
```@asset(freshness_policy=FreshnessPolicy(maximum_lag_minutes=20))```


ah yep that's right -- working on adding support for that either this week or next! the current plan is to add a new parameter to the `load_assets_from...` command, which will allow you to either supply a constant freshness policy (which will be applied to all loaded models), or a function that takes in a dictionary of node metadata and returns a freshness policy (i.e. `lambda node_info: FreshnessPolicy(maximum_lag_minutes=20) if node_info["name"] == "model_i_care_about" else None`)

<@U01J51Y6B9D> would this be the same for `load_assets_from_dbt_project` too?

Also, whilst waiting for this feature to be available in `load_assets_from...` , is there a way to do it by applying freshnessPolicy directly on the assets returned from `dbt_assets = load_assets_from_dbt_project(…)` ?

Heya, for anyone wondering, this feature has been implemented in dagster version `1.1.7` .

To enable freshness policy for a selected dbt model, you’ll just have to add the following line of config to your dbt model:

```# file: my_model.sql
{{
    config(
        dagster_freshness_policy={"maximum_lag_minutes": 60, "cron_schedule": "0 9 * * *"}
    )
}}
select * from {{ ref('my_table') }}```
More here: <https://github.com/dagster-io/dagster/issues/10924>