Skip to main content

Airflow Best Practices

This page should serve as a reference for tips and tricks that we recommend for the best Airflow experience. Please read the official Airflow Best Practices doc first.

Table of Contents​

This page aims to be a reference for airflow recommendations.

Start Date​

Do not use dynamic scheduled dates. Always set your start date for the day before or sooner and set catchup=false to avoid running additional runs:

from pendulum import datetime
from airflow.decorators import dag

@dag(
    default_args=("start_date": datetime(2023, 12, 29), # Set this to the day before or earlier
    "owner": "Noel Gomez",
    "email": "gomezn@example.com",
    "email_on_failure": True,
),
  dag_id="sample_dag",
  schedule="@daily",
  catchup=False, # Set this to false to avoid additional catchup runs
  tags=["version_1"],
  description="Datacoves Sample dag",
)
...