Skip to content
Tutorials Center
My learning

Topics

Lesson 14 of 37 ยท Scheduling and Dependencies

Airflow Schedules: Cron Expressions, Presets and Timetables

Learn how to schedule Airflow DAGs with cron expressions, presets such as @daily, timedelta intervals, None for manual runs and custom timetables, with time zone tips.

The schedule argument tells Airflow when to create DAG runs. It accepts several kinds of values: cron strings, presets, a timedelta, None, assets (events) and custom timetables. This lesson covers the time-based ones in detail, including time zones, which cause more production surprises than any other detail.

The schedules are Airflow arguments and read-only here. A runnable cron stand-in below lets you check what a cron expression means by listing the next times it fires.

Cron expressions

A cron expression has five fields:

TEXT
โ”Œโ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€ minute        (0-59)
โ”‚ โ”Œโ”€โ”€โ”€โ”€โ”€โ”€ hour          (0-23)
โ”‚ โ”‚ โ”Œโ”€โ”€โ”€โ”€ day of month  (1-31)
โ”‚ โ”‚ โ”‚ โ”Œโ”€โ”€ month         (1-12)
โ”‚ โ”‚ โ”‚ โ”‚ โ”Œ day of week   (0-6, Sunday = 0 or 7)
โ”‚ โ”‚ โ”‚ โ”‚ โ”‚
* * * * *

Examples:

Expression Meaning
0 6 * * * Every day at 06:00
*/15 * * * * Every 15 minutes
30 2 * * 1-5 02:30 on weekdays
0 0 1 * * Midnight on the first of each month
0 9 * * MON Mondays at 09:00
0 */4 * * * Every 4 hours, on the hour
Python
with DAG(dag_id="weekday_report", start_date=datetime(2025, 1, 1),
         schedule="30 2 * * 1-5", catchup=False) as dag:
    ...

Presets

Airflow understands short names:

Preset Equals Meaning
@once n/a Exactly one run
@continuous n/a A new run starts as soon as the last finishes
@hourly 0 * * * * Top of every hour
@daily 0 0 * * * Midnight every day
@weekly 0 0 * * 0 Midnight Sunday
@monthly 0 0 1 * * Midnight on the 1st
@yearly 0 0 1 1 * Midnight on 1 January

timedelta: fixed intervals

Python
from datetime import timedelta

with DAG(dag_id="every_90_minutes", start_date=datetime(2025, 1, 1),
         schedule=timedelta(minutes=90), catchup=False) as dag:
    ...

A timedelta produces runs spaced exactly that far apart, counted from start_date onward. Cron cannot express "every 90 minutes" because 90 does not divide a day evenly, so use timedelta (or a custom timetable) for such cases.

None: manual only

schedule=None means "never create runs automatically". Use it for DAGs you trigger by hand, from the API or from another DAG.

Cron semantics in Airflow 3

In Airflow 2, a cron string created data intervals: the run for the interval 1 March to 2 March started on 2 March at the end of the interval. In Airflow 3 a plain cron string uses the trigger timetable: a run is created at each cron time, with no interval attached. If you want the older behaviour, ask for it explicitly:

Python
from airflow.timetables.interval import CronDataIntervalTimetable

with DAG(
    dag_id="classic_interval",
    start_date=datetime(2025, 1, 1),
    schedule=CronDataIntervalTimetable("0 6 * * *", timezone="Europe/Oslo"),
) as dag:
    ...

and for trigger semantics with a time zone:

Python
from airflow.timetables.trigger import CronTriggerTimetable

schedule = CronTriggerTimetable("0 6 * * *", timezone="Europe/Oslo")

The next lesson explains intervals and the logical date in depth, because this choice decides which dates your task sees.

Time zones

By default Airflow works in UTC. Cron 0 6 * * * means 06:00 UTC. To schedule in local time, give the DAG a time-zone-aware start_date or build the timetable with timezone=:

Python
import pendulum

with DAG(
    dag_id="oslo_morning",
    start_date=pendulum.datetime(2025, 1, 1, tz="Europe/Oslo"),
    schedule="0 6 * * *",
    catchup=False,
) as dag:
    ...

Daylight saving time bites here. At the spring change a local 02:30 does not exist; at the autumn change 02:30 happens twice. Airflow's timetables handle this, but UTC is the simplest choice for data pipelines, and you can still display local time in the UI.

The first run and start_date

The earliest run is determined by start_date and the schedule. With catchup=False Airflow creates only the latest due run when the DAG is first unpaused (details in Catchup). Always use a fixed, past start_date.

Stand-in: next run times of a cron expression

This stand-in implements a small cron matcher for the five fields (*, numbers, ranges, lists and */n steps) and lists the next firing times from a start moment. It is a teaching tool, not Airflow's engine.

Python Stand-in: when does this cron fire?

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Try your own expression, for instance 0 9 * * MON will not work in this stand-in because names are not supported; use 0 9 * * 1.

Stand-in: timedelta schedule

A timedelta schedule is just repeated addition. This stand-in shows why 90 minutes cannot be a cron expression:

Python Stand-in: every 90 minutes

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The times drift across the day (00:00, 01:30, 03:00, ...) and the next day starts at 12:00, which a single cron line cannot produce.

Choosing a schedule

  • Business reports: cron at a fixed local time, with a time zone.
  • Pipelines chained to other DAGs: consider asset-based scheduling instead of guessing times.
  • "Every N units" that does not divide evenly: timedelta.
  • Ad-hoc jobs: None.
  • Streaming-like batches: @continuous with care.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Which cron expression runs at 02:30 on weekdays?

    Fields are minute, hour, day of month, month, day of week: 30 2 * * 1-5.
  2. 2. What does schedule=None mean?

    None disables automatic runs, which suits manual and externally triggered DAGs.
  3. 3. Why use timedelta(minutes=90) instead of cron for an every-90-minutes job?

    Cron fields repeat per hour/day boundary, so irregular periods need a fixed interval.
  4. 4. What time zone does Airflow use by default for schedules?

    UTC is the default; use a timezone-aware start_date or timetable for local time.

Summary

  • schedule accepts cron strings, presets, timedelta, None, timetables and assets.
  • Airflow 3 runs plain cron strings with trigger semantics; use CronDataIntervalTimetable for the classic interval behaviour.
  • Time zones default to UTC; mind daylight saving time if you schedule in local time.
  • Pick fixed past start_date values and think about whether events or time should drive each DAG.

Next: Data intervals and the logical date