How to create a graph_asset that calls the same op twice but with different configuration values for a ConfigurableResource?

[ukhsa-ri]

I'm trying to create a graph_asset that calls an op twice, this op calls a single ConfigurableResource but with different configuration values for the two calls. However I can't get this to work.

Intended use case: The single graph_asset represents a table where data is read in from 2 different servers/databases and joined together.

The below (simplified) works when the configuration for the resource is supplied to the op but this means the resource is set to one server/database only. I'd like to be able to reuse the op within the job to pull data from another server/database

from dagster import op
from sqlalchemy import create_engine
from sqlalchemy.engine import  URL
import dagster as dg
import dask.dataframe as dd
import pandas as pd


class ConnectionResource(dg.ConfigurableResource):
    server: str
    database: str
    
    def sql_engine(self, context: dg.InitResourceContext):
        ds = (
            r"DRIVER={ODBC Driver 17 for SQL Server};"
            f"SERVER={self.server};"  
            f"DATABASE={self.database};"  
            r"Trusted_Connection=yes;"
            r"TrustServerCertificate=yes;"
        )

        connection_url = URL.create("mssql+pyodbc", query={"odbc_connect": ds})
        engine = create_engine(connection_url, fast_executemany=True, future=True)       

        return engine

@op
def read_table_from_sql(
    context, sql: ConnectionResource, table_name: str
) -> dd.DataFrame:
    engine = sql.sql_engine(context)
    context.log.info(f"Reading table '{table_name}' from SQL Server")
    df = pd.read_sql_table(table_name, con=engine)
    df = dd.from_pandas(df)
    context.log.info(f"Read {len(df)} rows from '{table_name}'.")

    return df

@dg.graph_asset(
        config={
            "read_table_from_sql": {"inputs": {"table_name": "some_table_name"}},
        }
)
def asset_one():
    df = read_table_from_sql()

    # ideally including this 2nd read and merge but cannot configure a 2nd server/database using the same ConfigurableResource
    df2 = read_table_from_sql("but from a different server/database")
    df = df.merge(df2)

    return df


asset_job = dg.define_asset_job(
    name="asset_job",
    selection=[asset_one],
)

# Define your Definitions with resources
defs = dg.Definitions(
    assets=[asset_one],
    jobs=[asset_job],
    resources={
        "sql": ConnectionResource(
            server="someserver",
            database="somedatabase"
        )
    }
)

I have tried to supply resource_defs to the graph_asset but this leads to various issues and even Dagster AI doesn't seem to be able to suggest something that works?