Ok sorry. My bad, it is true that I set the

DAGSTER_HOME

environment variable just today, but this moring! Wait, this means that specifying the

DAGSTER_HOME

the storage folder with all result of materialization will be always available even offline without being removed from a temporary file? So it means that every time I will restart a Dagster service I will get all my asset materialized in previous session already materialized ?

Yes, setting the

DAGSTER_HOME

environment variable is an important step in ensuring that your Dagster instance has a persistent location to store metadata, including the results of asset materializations. When

DAGSTER_HOME

is set, Dagster uses the specified directory to store various instance-related data, such as run history, event logs, and output from materialized assets. If you do not set

DAGSTER_HOME

, Dagster will default to using a temporary directory for storage, which means that all data will be lost when the Dagster process exits. By setting

DAGSTER_HOME

to a specific directory on your filesystem, you ensure that the data is retained across Dagster service restarts, and you can access the results of previous materializations even when offline. It's important to note that the actual storage of materialized asset outputs is handled by the configured IO manager, which is responsible for writing outputs to and reading inputs from storage. If you are using the default filesystem IO manager and you have set

DAGSTER_HOME

, the outputs will be stored in a subdirectory within

DAGSTER_HOME

unless otherwise configured. In summary, setting

DAGSTER_HOME

allows you to have a consistent and persistent location for your Dagster instance's metadata and materialized asset data, which will be available across service restarts and offline access. DagsterInstance() on Dagster Cloud Serve... button [docs] - dagster.yaml file per project button Dagster Docs button Using environment variables and secrets ... button Thank you for your feedback!

for my DAGSTER_HOME variable I can set whatever folder I want? or I need to choose the one where my dagster code is located or the one pointed by my pyproject.toml