I got Airflow 2.8.2 installed with Docker.
I need clean Postgres database used with Airlow to free space on hard drive.

First of all this is my database memory usage:

As I know I can’t just delete rows from tables.
So I found simple dag for clearing database:

from airflow import settings
from airflow.models import DagRun
from sqlalchemy.orm import load_onlysession = settings.Session()def remove_old_dag_runs():
    query = session.query(
            DagRun # specify of which resource metadata should be removed
        ).options(
            load_only(
                DagRun.execution_date # just get the execution date column
            )
        )
    query = query.filter(
            DagRun.execution_date < '2022-02-25 12:23:38.462662+00'
        )
    query.delete(synchronize_session=False)
    session.commit()remove_old_dag_runs_task = PythonOperator(
    task_id='remove_old_dag_runs',
    python_callable=remove_old_dag_runs,
    dag=dag
)

Also I found this code for DB cleaning.

But this code used with old Airflow version.

So which database object I can delete in Airflow 2.8.2?
I got this list of objects:

airflow.models list 1:
DagModel, DagRun, Log, XCom, SlaMiss, TaskInstance, Variable
airflow.models list 2:
TaskReschedule, TaskFail, RenderedTaskInstanceFields, ImportError

airflow.jobs list 1:
from airflow.jobs.base_job import BaseJob
airflow.jobs list 2:
from airflow.jobs.scheduler_job import SchedulerJob
from airflow.jobs.local_task_job import LocalTaskJob