Schedules and Sensors

Jobs define which assets to run. Schedules and sensors define when to run them.

Schedules

A schedule triggers jobs on a time-based interval using cron notation:

from pathlib import Path
import polars as pl
import dagster as dg
from dagster_duckdb_polars import DuckDBPolarsIOManager


@dg.asset(io_manager_key="io_manager")
def plays() -> pl.DataFrame:
    date = Path("2025/10/16")
    df = pl.read_csv(date / "plays.csv")
    return df


io_manager = DuckDBPolarsIOManager(database="db.duckdb")
plays_job = dg.define_asset_job("plays_job", selection=["plays"])

daily_schedule = dg.ScheduleDefinition(
    name="daily_refresh",
    cron_schedule="0 0 * * *",  # Runs at midnight daily
    job=plays_job,
    default_status=dg.DefaultScheduleStatus.RUNNING,
)


def definitions() -> dg.Definitions:
    return dg.Definitions(
        assets=[plays],
        resources={"io_manager": io_manager},
        jobs=[plays_job],
        schedules=[daily_schedule],
    )

Setting default_status=RUNNING enables the schedule immediately. Otherwise, it must be enabled through the UI or CLI.

Sensors

Sensors monitor external events and trigger jobs when conditions are met. Unlike schedules, sensors are event-driven rather than time-driven.

Example use cases:

• Trigger when a new file appears in S3
• Trigger when a database table is updated
• Trigger when an external API returns new data

Sensors poll at a configurable interval and evaluate whether to trigger a run based on custom logic.