Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Order triggers by - TI priority_weight when assign unassigned triggers #32318

Merged
merged 6 commits into from Aug 14, 2023
Merged

Order triggers by - TI priority_weight when assign unassigned triggers #32318

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
a3bacdb
Order triggers by - TI priority_weight when assign unassigned triggers
hussein-awala Jul 2, 2023
c266d1c
Update airflow/models/trigger.py
hussein-awala Jul 10, 2023
b226ce5
Merge branch 'main' into sorted_triggers
eladkal Jul 20, 2023
45fa69b
Merge branch 'main' into sorted_triggers
hussein-awala Aug 10, 2023
b0fe3b8
Replace outer join by inner join and use coalesce to handle None values
hussein-awala Aug 10, 2023
9fa2ed4
fix unit tests
hussein-awala Aug 11, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
3 changes: 2 additions & 1 deletion airflow/models/trigger.py
View file
Open in desktop
Expand Up @@ -241,8 +241,9 @@ def assign_unassigned(cls, triggerer_id, capacity, heartrate, session: Session =
def get_sorted_triggers(cls, capacity, alive_triggerer_ids, session):
return with_row_locks(
session.query(cls.id)
.join(TaskInstance, cls.id == TaskInstance.trigger_id, isouter=True)
.filter(or_(cls.triggerer_id.is_(None), cls.triggerer_id.notin_(alive_triggerer_ids)))
.order_by(cls.created_date)
.order_by(-TaskInstance.priority_weight, cls.created_date)
Copy link
Member

@uranusjr uranusjr Jul 3, 2023 •
edited

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
.order_by(-TaskInstance.priority_weight, cls.created_date)
.order_by(TaskInstance.priority_weight.desc(), cls.created_date)

Preferred style.

Can this be NULL and how do we want to handle them?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For NULL, I can use TaskInstance.priority_weight.desc().nulls_last(), but it seems like the lock (with_for_update) doesn't work with the outer join, I will try to find a solution for that.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we use CASE instead? We don’t really need to put them last, just probably not at the very beginning.

.limit(capacity),
session,
skip_locked=True,
Expand Down
88 changes: 79 additions & 9 deletions tests/models/test_trigger.py
View file
Open in desktop
Expand Up @@ -243,31 +243,101 @@ def test_assign_unassigned_missing_heartbeat(session, create_task_instance, chec
second_triggerer.latest_heartbeat += datetime.timedelta(seconds=check_triggerer_heartrate)


def test_get_sorted_triggers(session, create_task_instance):
def test_get_sorted_triggers_same_priority_weight(session, create_task_instance):
"""
Tests that triggers are sorted by the creation_date.
Tests that triggers are sorted by the creation_date if they have the same priority.
"""
old_execution_date = datetime.datetime(
2023, 5, 9, 12, 16, 14, 474415, tzinfo=pytz.timezone("Africa/Abidjan")
)
trigger_old = Trigger(
classpath="airflow.triggers.testing.SuccessTrigger",
kwargs={},
created_date=datetime.datetime(
2023, 5, 9, 12, 16, 14, 474415, tzinfo=pytz.timezone("Africa/Abidjan")
),
created_date=old_execution_date + datetime.timedelta(seconds=30),
)
trigger_old.id = 1
session.add(trigger_old)
TI_old = create_task_instance(
task_id="old",
execution_date=old_execution_date,
run_id="old_run_id",
)
TI_old.priority_weight = 1
TI_old.trigger_id = trigger_old.id
session.add(TI_old)

new_execution_date = datetime.datetime(
2023, 5, 9, 12, 17, 14, 474415, tzinfo=pytz.timezone("Africa/Abidjan")
)
trigger_new = Trigger(
classpath="airflow.triggers.testing.SuccessTrigger",
kwargs={},
created_date=datetime.datetime(
2023, 5, 9, 12, 17, 14, 474415, tzinfo=pytz.timezone("Africa/Abidjan")
),
created_date=new_execution_date + datetime.timedelta(seconds=30),
)
trigger_new.id = 2
session.add(trigger_old)
session.add(trigger_new)
TI_new = create_task_instance(
task_id="new",
execution_date=new_execution_date,
run_id="new_run_id",
)
TI_new.priority_weight = 1
TI_new.trigger_id = trigger_new.id
session.add(TI_new)

session.commit()
assert session.query(Trigger).count() == 2

trigger_ids_query = Trigger.get_sorted_triggers(capacity=100, alive_triggerer_ids=[], session=session)

assert trigger_ids_query == [(1,), (2,)]


def test_get_sorted_triggers_different_priority_weights(session, create_task_instance):
"""
Tests that triggers are sorted by the priority_weight.
"""
old_execution_date = datetime.datetime(
2023, 5, 9, 12, 16, 14, 474415, tzinfo=pytz.timezone("Africa/Abidjan")
)
trigger_old = Trigger(
classpath="airflow.triggers.testing.SuccessTrigger",
kwargs={},
created_date=old_execution_date + datetime.timedelta(seconds=30),
)
trigger_old.id = 1
session.add(trigger_old)
TI_old = create_task_instance(
task_id="old",
execution_date=old_execution_date,
run_id="old_run_id",
)
TI_old.priority_weight = 1
TI_old.trigger_id = trigger_old.id
session.add(TI_old)

new_execution_date = datetime.datetime(
2023, 5, 9, 12, 17, 14, 474415, tzinfo=pytz.timezone("Africa/Abidjan")
)
trigger_new = Trigger(
classpath="airflow.triggers.testing.SuccessTrigger",
kwargs={},
created_date=new_execution_date + datetime.timedelta(seconds=30),
)
trigger_new.id = 2
session.add(trigger_new)
TI_new = create_task_instance(
task_id="new",
execution_date=new_execution_date,
run_id="new_run_id",
)
TI_new.priority_weight = 2
TI_new.trigger_id = trigger_new.id
session.add(TI_new)

session.commit()
assert session.query(Trigger).count() == 2

trigger_ids_query = Trigger.get_sorted_triggers(capacity=100, alive_triggerer_ids=[], session=session)

assert trigger_ids_query == [(2,), (1,)]