airflow-systemd 0.1.0

Systemd operators and configuration for long-running tasks

airflow-systemd

Systemd operators and configuration for long-running tasks

Overview

airflow-systemd mirrors airflow-supervisor for hosts managed by systemd. It turns a SystemdAirflowConfiguration into an Airflow task lifecycle:

configure-systemd -> start-services -> check-services
                                      |             |
                                      v             v
                               restart-services  stop-services -> unconfigure-systemd

Airflow owns scheduling, so this integration runs service units directly rather than starting systemd timer units. Timer models remain available through systemd-pydantic for non-Airflow consumers.

from airflow import DAG
from airflow_systemd import ServiceConfiguration, ServiceUnitConfiguration, Systemd, SystemdAirflowConfiguration

dag = DAG(dag_id="long-running-job", schedule="@daily")
config = SystemdAirflowConfiguration(
    service={
        "long-running-job": ServiceUnitConfiguration(
            service=ServiceConfiguration(type="exec", exec_start="/opt/jobs/run"),
        )
    },
    scope="user",
)
Systemd(dag=dag, cfg=config)

The local integration supports the same Airflow-specific controls as airflow-supervisor: status interval and timeout, runtime/end-time limits, retrigger limits, pools, optional restart behavior, stop-on-exit, and cleanup.

[!NOTE] This library was generated using copier from the Base Python Project Template repository.