Merge pull request #3437 from alphagov/retry-parallel-status-180693991

Attempt 2 of parallelising status aggregation more
This commit is contained in:
Ben Thorner
2022-02-14 10:50:26 +00:00
committed by GitHub
9 changed files with 166 additions and 87 deletions
+2 -2
View File
@@ -22,7 +22,7 @@ from app.dao.jobs_dao import (
from app.dao.notifications_dao import (
dao_get_notifications_processing_time_stats,
dao_timeout_notifications,
get_service_ids_that_have_notifications_from_before_timestamp,
get_service_ids_with_notifications_before,
move_notifications_to_notification_history,
)
from app.dao.service_data_retention_dao import (
@@ -107,7 +107,7 @@ def _delete_notifications_older_than_retention_by_type(notification_type):
# get a list of all service ids that we'll need to delete for. Typically that might only be 5% of services.
# This query takes a couple of mins to run.
service_ids_that_have_sent_notifications_recently = get_service_ids_that_have_notifications_from_before_timestamp(
service_ids_that_have_sent_notifications_recently = get_service_ids_with_notifications_before(
notification_type,
seven_days_ago
)
+28 -29
View File
@@ -14,7 +14,8 @@ from app.dao.fact_notification_status_dao import (
fetch_status_data_for_service_and_day,
update_fact_notification_status,
)
from app.models import EMAIL_TYPE, LETTER_TYPE, SMS_TYPE, Service
from app.dao.notifications_dao import get_service_ids_with_notifications_on_date
from app.models import EMAIL_TYPE, LETTER_TYPE, SMS_TYPE
@notify_celery.task(name="create-nightly-billing")
@@ -97,52 +98,50 @@ def create_nightly_notification_status():
for i in range(days):
process_day = yesterday - timedelta(days=i)
create_nightly_notification_status_for_day.apply_async(
kwargs={
'process_day': process_day.isoformat(),
'notification_type': notification_type,
},
queue=QueueNames.REPORTING
)
current_app.logger.info(
f"create-nightly-notification-status-for-day task created "
f"for {notification_type} and {process_day}"
relevant_service_ids = get_service_ids_with_notifications_on_date(
notification_type, process_day
)
for service_id in relevant_service_ids:
create_nightly_notification_status_for_service_and_day.apply_async(
kwargs={
'process_day': process_day.isoformat(),
'notification_type': notification_type,
'service_id': service_id,
},
queue=QueueNames.REPORTING
)
@notify_celery.task(name="create-nightly-notification-status-for-day")
def create_nightly_notification_status_for_day(process_day, notification_type):
@notify_celery.task(name="create-nightly-notification-status-for-service-and-day")
def create_nightly_notification_status_for_service_and_day(process_day, service_id, notification_type):
process_day = datetime.strptime(process_day, "%Y-%m-%d").date()
current_app.logger.info(
f'create-nightly-notification-status-for-day task started '
f'for {notification_type} for {process_day}'
)
start = datetime.utcnow()
new_status_rows = []
for service in Service.query.all():
new_status_rows += fetch_status_data_for_service_and_day(
process_day=process_day,
notification_type=notification_type,
service_id=service.id,
)
new_status_rows = fetch_status_data_for_service_and_day(
process_day=process_day,
notification_type=notification_type,
service_id=service_id,
)
end = datetime.utcnow()
current_app.logger.info(
f'create-nightly-notification-status-for-day task fetch '
f'for {notification_type} for {process_day}: '
f'for {service_id}, {notification_type} for {process_day}: '
f'data fetched in {(end - start).seconds} seconds'
)
start = datetime.utcnow()
update_fact_notification_status(
new_status_rows=new_status_rows,
process_day=process_day,
notification_type=notification_type,
service_id=service_id
)
end = datetime.utcnow()
current_app.logger.info(
f'create-nightly-notification-status-for-day task finished '
f'for {notification_type} for {process_day}: '
f'{len(new_status_rows)} rows updated'
f'create-nightly-notification-status-for-day task update '
f'for {service_id}, {notification_type} for {process_day}: '
f'data fetched in {(end - start).seconds} seconds'
)
+4 -4
View File
@@ -1,7 +1,7 @@
from datetime import date, datetime, time, timedelta
from datetime import date, datetime, timedelta
from flask import current_app
from notifications_utils.timezones import convert_bst_to_utc, convert_utc_to_bst
from notifications_utils.timezones import convert_utc_to_bst
from sqlalchemy import Date, Integer, and_, desc, func
from sqlalchemy.dialects.postgresql import insert
from sqlalchemy.sql.expression import case, literal
@@ -317,8 +317,8 @@ def delete_billing_data_for_service_for_day(process_day, service_id):
def fetch_billing_data_for_day(process_day, service_id=None, check_permissions=False):
start_date = convert_bst_to_utc(datetime.combine(process_day, time.min))
end_date = convert_bst_to_utc(datetime.combine(process_day + timedelta(days=1), time.min))
start_date = get_london_midnight_in_utc(process_day)
end_date = get_london_midnight_in_utc(process_day + timedelta(days=1))
current_app.logger.info("Populate ft_billing for {} to {}".format(start_date, end_date))
transit_data = []
if not service_id:
+6 -8
View File
@@ -1,6 +1,5 @@
from datetime import datetime, time, timedelta
from datetime import datetime, timedelta
from notifications_utils.timezones import convert_bst_to_utc
from sqlalchemy import Date, case, func
from sqlalchemy.dialects.postgresql import insert
from sqlalchemy.sql.expression import extract, literal
@@ -36,8 +35,8 @@ from app.utils import (
def fetch_status_data_for_service_and_day(process_day, service_id, notification_type):
start_date = convert_bst_to_utc(datetime.combine(process_day, time.min))
end_date = convert_bst_to_utc(datetime.combine(process_day + timedelta(days=1), time.min))
start_date = get_london_midnight_in_utc(process_day)
end_date = get_london_midnight_in_utc(process_day + timedelta(days=1))
# query notifications or notification_history for the day, depending on their data retention
service = Service.query.get(service_id)
@@ -45,7 +44,6 @@ def fetch_status_data_for_service_and_day(process_day, service_id, notification_
return db.session.query(
table.template_id,
table.service_id,
func.coalesce(table.job_id, '00000000-0000-0000-0000-000000000000').label('job_id'),
table.key_type,
table.status,
@@ -58,7 +56,6 @@ def fetch_status_data_for_service_and_day(process_day, service_id, notification_
table.key_type.in_((KEY_TYPE_NORMAL, KEY_TYPE_TEAM)),
).group_by(
table.template_id,
table.service_id,
'job_id',
table.key_type,
table.status
@@ -66,11 +63,12 @@ def fetch_status_data_for_service_and_day(process_day, service_id, notification_
@autocommit
def update_fact_notification_status(new_status_rows, process_day, notification_type):
def update_fact_notification_status(new_status_rows, process_day, notification_type, service_id):
table = FactNotificationStatus.__table__
FactNotificationStatus.query.filter(
FactNotificationStatus.bst_date == process_day,
FactNotificationStatus.notification_type == notification_type,
FactNotificationStatus.service_id == service_id,
).delete()
for row in new_status_rows:
@@ -78,7 +76,7 @@ def update_fact_notification_status(new_status_rows, process_day, notification_t
insert(table).values(
bst_date=process_day,
template_id=row.template_id,
service_id=row.service_id,
service_id=service_id,
job_id=row.job_id,
notification_type=notification_type,
key_type=row.key_type,
+23 -2
View File
@@ -46,7 +46,11 @@ from app.models import (
NotificationHistory,
ProviderDetails,
)
from app.utils import escape_special_characters, midnight_n_days_ago
from app.utils import (
escape_special_characters,
get_london_midnight_in_utc,
midnight_n_days_ago,
)
def dao_get_last_date_template_was_used(template_id, service_id):
@@ -784,7 +788,7 @@ def _duplicate_update_warning(notification, status):
)
def get_service_ids_that_have_notifications_from_before_timestamp(notification_type, timestamp):
def get_service_ids_with_notifications_before(notification_type, timestamp):
return {
row.service_id
for row in db.session.query(
@@ -794,3 +798,20 @@ def get_service_ids_that_have_notifications_from_before_timestamp(notification_t
Notification.created_at < timestamp
).distinct()
}
def get_service_ids_with_notifications_on_date(notification_type, date):
start_date = get_london_midnight_in_utc(date)
end_date = get_london_midnight_in_utc(date + timedelta(days=1))
return {
row.service_id
for row in db.session.query(
Notification.service_id
).filter(
Notification.notification_type == notification_type,
# using >= + < is much more efficient than date(created_at)
Notification.created_at >= start_date,
Notification.created_at < end_date,
).distinct()
}
+2 -4
View File
@@ -8,7 +8,7 @@ from notifications_utils.template import (
LetterPrintTemplate,
SMSMessageTemplate,
)
from notifications_utils.timezones import convert_utc_to_bst
from notifications_utils.timezones import convert_bst_to_utc, convert_utc_to_bst
from sqlalchemy import func
DATETIME_FORMAT_NO_TIMEZONE = "%Y-%m-%d %H:%M:%S.%f"
@@ -64,9 +64,7 @@ def get_london_midnight_in_utc(date):
:param date: the day to calculate the London midnight in UTC for
:return: the datetime of London midnight in UTC, for example 2016-06-17 = 2016-06-16 23:00:00
"""
return local_timezone.localize(datetime.combine(date, datetime.min.time())).astimezone(
pytz.UTC).replace(
tzinfo=None)
return convert_bst_to_utc(datetime.combine(date, datetime.min.time()))
def get_midnight_for_day_before(date):
+58 -37
View File
@@ -1,4 +1,3 @@
import itertools
from datetime import date, datetime, time, timedelta
from decimal import Decimal
from uuid import UUID
@@ -10,7 +9,7 @@ from app.celery.reporting_tasks import (
create_nightly_billing,
create_nightly_billing_for_day,
create_nightly_notification_status,
create_nightly_notification_status_for_day,
create_nightly_notification_status_for_service_and_day,
)
from app.config import QueueNames
from app.dao.fact_billing_dao import get_rate
@@ -20,6 +19,7 @@ from app.models import (
KEY_TYPE_TEAM,
KEY_TYPE_TEST,
LETTER_TYPE,
NOTIFICATION_TYPES,
SMS_TYPE,
FactBilling,
FactNotificationStatus,
@@ -61,37 +61,58 @@ def test_create_nightly_billing_triggers_tasks_for_days(notify_api, mocker, day_
assert mock_celery.apply_async.call_args_list[i][1]['kwargs'] == {'process_day': expected_kwargs[i]}
@freeze_time('2019-08-01')
def test_create_nightly_notification_status_triggers_tasks(notify_api, sample_service, mocker):
mock_celery = mocker.patch('app.celery.reporting_tasks.create_nightly_notification_status_for_day')
@freeze_time('2019-08-01T00:30')
def test_create_nightly_notification_status_triggers_tasks(
notify_api,
sample_service,
sample_template,
mocker,
):
mock_celery = mocker.patch(
'app.celery.reporting_tasks.create_nightly_notification_status_for_service_and_day'
).apply_async
create_notification(template=sample_template, created_at='2019-07-31')
create_nightly_notification_status()
assert mock_celery.apply_async.call_count == (
(4 * 3) # four days, three notification types
+
6 # six more days of just letters
mock_celery.assert_called_with(
kwargs={
'service_id': sample_service.id,
'process_day': '2019-07-31',
'notification_type': SMS_TYPE
},
queue=QueueNames.REPORTING
)
for process_date, notification_type in itertools.product(
['2019-07-31', '2019-07-30', '2019-07-29', '2019-07-28'],
[SMS_TYPE, EMAIL_TYPE, LETTER_TYPE]
):
mock_celery.apply_async.assert_any_call(
kwargs={
'process_day': process_date,
'notification_type': notification_type,
},
queue=QueueNames.REPORTING
)
for process_date in ['2019-07-27', '2019-07-26', '2019-07-25', '2019-07-24', '2019-07-23', '2019-07-22']:
mock_celery.apply_async.assert_any_call(
kwargs={
'process_day': process_date,
'notification_type': LETTER_TYPE,
},
queue=QueueNames.REPORTING
)
@freeze_time('2019-08-01T00:30')
@pytest.mark.parametrize('notification_date, expected_types_aggregated', [
('2019-08-01', set()),
('2019-07-31', {EMAIL_TYPE, SMS_TYPE, LETTER_TYPE}),
('2019-07-28', {EMAIL_TYPE, SMS_TYPE, LETTER_TYPE}),
('2019-07-27', {LETTER_TYPE}),
('2019-07-22', {LETTER_TYPE}),
('2019-07-21', set()),
])
def test_create_nightly_notification_status_triggers_relevant_tasks(
notify_api,
sample_service,
mocker,
notification_date,
expected_types_aggregated,
):
mock_celery = mocker.patch(
'app.celery.reporting_tasks.create_nightly_notification_status_for_service_and_day'
).apply_async
for notification_type in NOTIFICATION_TYPES:
template = create_template(sample_service, template_type=notification_type)
create_notification(template=template, created_at=notification_date)
create_nightly_notification_status()
types = {call.kwargs['kwargs']['notification_type'] for call in mock_celery.mock_calls}
assert types == expected_types_aggregated
@pytest.mark.parametrize('second_rate, records_num, billable_units, multiplier',
@@ -507,7 +528,7 @@ def test_create_nightly_billing_for_day_update_when_record_exists(
assert records[0].updated_at
def test_create_nightly_notification_status_for_day(notify_db_session):
def test_create_nightly_notification_status_for_service_and_day(notify_db_session):
first_service = create_service(service_name='First Service')
first_template = create_template(service=first_service)
second_service = create_service(service_name='second Service')
@@ -538,9 +559,9 @@ def test_create_nightly_notification_status_for_day(notify_db_session):
assert len(FactNotificationStatus.query.all()) == 0
create_nightly_notification_status_for_day(str(process_day), 'sms')
create_nightly_notification_status_for_day(str(process_day), 'email')
create_nightly_notification_status_for_day(str(process_day), 'letter')
create_nightly_notification_status_for_service_and_day(str(process_day), first_service.id, 'sms')
create_nightly_notification_status_for_service_and_day(str(process_day), second_service.id, 'email')
create_nightly_notification_status_for_service_and_day(str(process_day), third_service.id, 'letter')
new_fact_data = FactNotificationStatus.query.order_by(
FactNotificationStatus.notification_type
@@ -575,13 +596,13 @@ def test_create_nightly_notification_status_for_day(notify_db_session):
assert new_fact_data[2].key_type == KEY_TYPE_NORMAL
def test_create_nightly_notification_status_for_day_overwrites_old_data(notify_db_session):
def test_create_nightly_notification_status_for_service_and_day_overwrites_old_data(notify_db_session):
first_service = create_service(service_name='First Service')
first_template = create_template(service=first_service)
create_notification(template=first_template, status='delivered')
process_day = date.today()
create_nightly_notification_status_for_day(str(process_day), 'sms')
create_nightly_notification_status_for_service_and_day(str(process_day), first_service.id, 'sms')
new_fact_data = FactNotificationStatus.query.order_by(
FactNotificationStatus.bst_date,
@@ -592,7 +613,7 @@ def test_create_nightly_notification_status_for_day_overwrites_old_data(notify_d
assert new_fact_data[0].notification_count == 1
create_notification(template=first_template, status='delivered')
create_nightly_notification_status_for_day(str(process_day), 'sms')
create_nightly_notification_status_for_service_and_day(str(process_day), first_service.id, 'sms')
updated_fact_data = FactNotificationStatus.query.order_by(
FactNotificationStatus.bst_date,
@@ -605,7 +626,7 @@ def test_create_nightly_notification_status_for_day_overwrites_old_data(notify_d
# the job runs at 12:30am London time. 04/01 is in BST.
@freeze_time('2019-04-01T23:30')
def test_create_nightly_notification_status_for_day_respects_bst(sample_template):
def test_create_nightly_notification_status_for_service_and_day_respects_bst(sample_template):
create_notification(sample_template, status='delivered', created_at=datetime(2019, 4, 1, 23, 0)) # too new
create_notification(sample_template, status='created', created_at=datetime(2019, 4, 1, 22, 59))
@@ -613,7 +634,7 @@ def test_create_nightly_notification_status_for_day_respects_bst(sample_template
create_notification(sample_template, status='delivered', created_at=datetime(2019, 3, 31, 22, 59)) # too old
create_nightly_notification_status_for_day('2019-04-01', 'sms')
create_nightly_notification_status_for_service_and_day('2019-04-01', sample_template.service_id, 'sms')
noti_status = FactNotificationStatus.query.order_by(FactNotificationStatus.bst_date).all()
assert len(noti_status) == 1
@@ -1,5 +1,5 @@
import uuid
from datetime import datetime, timedelta
from datetime import date, datetime, timedelta
from functools import partial
import pytest
@@ -24,6 +24,7 @@ from app.dao.notifications_dao import (
get_notification_with_personalisation,
get_notifications_for_job,
get_notifications_for_service,
get_service_ids_with_notifications_on_date,
is_delivery_slow_for_providers,
notifications_not_yet_sent,
update_notification_status_by_id,
@@ -41,6 +42,7 @@ from app.models import (
NOTIFICATION_STATUS_TYPES,
NOTIFICATION_STATUS_TYPES_FAILED,
NOTIFICATION_TEMPORARY_FAILURE,
SMS_TYPE,
Job,
Notification,
NotificationHistory,
@@ -1721,3 +1723,22 @@ def test_dao_get_letters_and_sheets_volume_by_postage(notify_db_session):
for result in results:
assert result._asdict() in expected_results
@pytest.mark.parametrize('created_at_utc,date_to_check,expected_count', [
# Clocks change on the 27th of March 2022, so the query needs to look at the
# time range 00:00 - 23:00 (UTC) thereafter.
('2022-03-27T00:30', date(2022, 3, 27), 1), # 27/03 00:30 GMT
('2022-03-27T22:30', date(2022, 3, 27), 1), # 27/03 23:30 BST
('2022-03-27T23:30', date(2022, 3, 27), 0), # 28/03 00:30 BST
('2022-03-26T23:30', date(2022, 3, 26), 1), # 26/03 23:30 GMT
])
def test_get_service_ids_with_notifications_on_date_respects_gmt_bst(
sample_template,
created_at_utc,
date_to_check,
expected_count
):
create_notification(template=sample_template, created_at=created_at_utc)
service_ids = get_service_ids_with_notifications_on_date(SMS_TYPE, date_to_check)
assert len(service_ids) == expected_count
@@ -14,6 +14,7 @@ from app.dao.fact_notification_status_dao import (
fetch_notification_status_totals_for_all_services,
fetch_notification_statuses_for_job,
fetch_stats_for_all_services_by_date_range,
fetch_status_data_for_service_and_day,
get_total_notifications_for_date_range,
)
from app.models import (
@@ -607,3 +608,23 @@ def test_get_total_notifications_for_date_range(sample_service):
assert len(results) == 1
assert results[0] == ("2021-03-01", 15, 20, 3)
@pytest.mark.parametrize('created_at_utc,process_day,expected_count', [
# Clocks change on the 27th of March 2022, so the query needs to look at the
# time range 00:00 - 23:00 (UTC) thereafter.
('2022-03-27T00:30', date(2022, 3, 27), 1), # 27/03 00:30 GMT
('2022-03-27T22:30', date(2022, 3, 27), 1), # 27/03 23:30 BST
('2022-03-27T23:30', date(2022, 3, 27), 0), # 28/03 00:30 BST
('2022-03-26T23:30', date(2022, 3, 26), 1), # 26/03 23:30 GMT
])
def test_fetch_status_data_for_service_and_day_respects_gmt_bst(
sample_template,
sample_service,
created_at_utc,
process_day,
expected_count,
):
create_notification(template=sample_template, created_at=created_at_utc)
rows = fetch_status_data_for_service_and_day(process_day, sample_service.id, SMS_TYPE)
assert len(rows) == expected_count