Airflow Usage Minutes

README.md

Dagster Cloud pricing is based on usage. If you are using Airflow, you can get an estimate of this usage by querying the Airflow metadata database.

For SQLite:

# navigate to your airflow DB, normally ~/airflow
sqlite3
.open airflowdb
SELECT SUM((julianday(end_date) - julianday(start_date))* 24 * 60) AS run_minutes FROM dag_run;

The Airflow dag_run table has the following fields which you can use to group or filter this query:

dag_id
run_id
state
run_type

For Postgres the SQL is similar. This query breaks down the run usage by month:

SELECT 
  date_trunc('month', start_date) AS month, 
  SUM(EXTRACT(epoch FROM (end_date - start_date))/60) AS run_minutes 
FROM dag_run 
GROUP BY 1;