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

dagster

how is one supposed to choose an io_manager for an asset defined from a graph?

i have a graph:
```@graph(tags={"notify": "slack"})
def splunk_augmentation_graph():
    xxx = start_splunk_run()
    return metadata_logic(
        RUN_ID,
        new_ips_df,
        get_ssl_cert_info(xxx,
        get_clearbit_reveal(xxx),
        get_registry_info(xxx),
        get_nslookup_info(xxx),
    )```

```splunk_metadata_dev = AssetsDefinition.from_graph(
    splunk_augmentation_graph,
    partitions_def=splunk_dev_schedule,
    keys_by_output_name={
        "result": AssetKey(["augmentation_data", "splunk_metadata"])
    },
)

splunk_metadata_group_dev = with_resources(
    [splunk_metadata_dev],
    resource_defs=RESOURCES,
)
splunk_metadata_dev_job = define_asset_job(
    name="splunk_metadata_dev_job",
    selection=AssetSelection.assets(splunk_metadata_dev),
    partitions_def=splunk_dev_schedule,
)```

the metadata_logic op is the only place i see to set an io_manager.

```@op(
    out={
        "splunk_metadata": Out(
            asset_key=AssetKey(["augmentation_data", "splunk_metadata"]),
            metadata={"partition_expr": "created_on"},
            io_manager_key="snowflake_io_manager",
        )
    }
)
def metadata_logic(```

You can provide the io manager in `with_resources` through `resource_defs`. Here’s an example: <https://docs.dagster.io/concepts/io-management/io-managers#applying-io-managers-to-assets>

i had the out definition messed up. got it fixed.