Airflow, Connecting to MsSql error "Adaptive Server is unavailable or does not exist" - airflow

Question

I'm getting this error when trying to use Airflow to get_records().
pymssql.OperationalError: (20009, b'DB-Lib error message 20009, severity 9:\nUnable to connect: Adaptive Server is unavailable or does not exist (localhost:None)\n')
I used this guide on how to setup.
https://tech.marksblogg.com/mssql-sql-server-linux-install-tutorial-and-guide.html
Using Python REPL, I can connect and return a result.
with pymssql.connect(server="localhost",
user="SA",
password="password",
database="database_name") as conn:
df = pd.read_sql("SELECT TOP 1 * FROM currency", conn)
print(df)
with pymssql.connect(server="127.0.0.1",
user="SA",
password="password",
database="database_name") as conn:
df = pd.read_sql("SELECT TOP 1 * FROM currency", conn)
print(df)
I update my Airflow Connections with either of these setups and the run a test.
airflow test run_test_db test_database 2015-06-01
The error is produced....
Any ideas please? The whole setup is contained within the one linux (vagrant) machine, no virtual environments.... So its using the same version of pymssql to try an connect....
EDIT UPDATE
Whats really annoying is if i use the same connection string in a DAG there is no error and it runs perfectly fine...
So the connection string that is retrieved from the database must change.
Is there a way to debug/print the string/connection properties?
Example working DAG
from airflow import DAG
from airflow.operators.python_operator import PythonOperator
from airflow.hooks.mssql_hook import MsSqlHook
from datetime import datetime, timedelta
import pymssql
import pandas as pd
default_args = {
'owner': 'airflow',
'depends_on_past': True,
'start_date': datetime(2019, 2, 13),
'email': ['example#email.com'],
'email_on_failure': False,
'email_on_retry': False,
'retries': 1,
'retry_delay': timedelta(minutes=5),
}
dag = DAG('run_test_db', default_args=default_args, schedule_interval="0 01 * * 2-6")
def test_db(**context):
with pymssql.connect(server="localhost",
user="SA",
password="Password123",
database="database") as conn:
df = pd.read_sql("SELECT TOP 1 * FROM champ_dw_dim_currency", conn)
print(df)
test_database = PythonOperator(
task_id='test_database',
python_callable=test_db,
execution_timeout=timedelta(minutes=3),
dag=dag,
provide_context=True,
op_kwargs={
'extra_detail': 'nothing'
})