Rewrite task to aggregate status by service

This is a step towards parallelising the task by service and day.
This commit is contained in:
Ben Thorner
2022-01-11 16:33:56 +00:00
parent 9fc8b904c6
commit ddbf556486
2 changed files with 42 additions and 39 deletions

View File

@@ -3,7 +3,7 @@ from datetime import datetime, timedelta
from flask import current_app from flask import current_app
from notifications_utils.timezones import convert_utc_to_bst from notifications_utils.timezones import convert_utc_to_bst
from app import notify_celery from app import db, notify_celery
from app.config import QueueNames from app.config import QueueNames
from app.cronitor import cronitor from app.cronitor import cronitor
from app.dao.fact_billing_dao import ( from app.dao.fact_billing_dao import (
@@ -11,10 +11,10 @@ from app.dao.fact_billing_dao import (
update_fact_billing, update_fact_billing,
) )
from app.dao.fact_notification_status_dao import ( from app.dao.fact_notification_status_dao import (
fetch_notification_status_for_day, fetch_status_data_for_service_and_day,
update_fact_notification_status, update_fact_notification_status,
) )
from app.models import EMAIL_TYPE, LETTER_TYPE, SMS_TYPE from app.models import EMAIL_TYPE, LETTER_TYPE, SMS_TYPE, Service
@notify_celery.task(name="create-nightly-billing") @notify_celery.task(name="create-nightly-billing")
@@ -124,15 +124,26 @@ def create_nightly_notification_status_for_day(process_day, notification_type):
f'create-nightly-notification-status-for-day task for {process_day} type {notification_type}: started' f'create-nightly-notification-status-for-day task for {process_day} type {notification_type}: started'
) )
for (service_id,) in db.session.query(Service.id):
start = datetime.utcnow() start = datetime.utcnow()
transit_data = fetch_notification_status_for_day(process_day=process_day, notification_type=notification_type) transit_data = fetch_status_data_for_service_and_day(
process_day=process_day,
notification_type=notification_type,
service_id=service_id,
)
end = datetime.utcnow() end = datetime.utcnow()
current_app.logger.info( current_app.logger.info(
f'create-nightly-notification-status-for-day task for {process_day} type {notification_type}: ' f'create-nightly-notification-status-for-day task for {process_day} type {notification_type}: '
f'data fetched in {(end - start).seconds} seconds' f'data fetched in {(end - start).seconds} seconds'
) )
update_fact_notification_status(transit_data, process_day, notification_type) update_fact_notification_status(
transit_data=transit_data,
process_day=process_day,
notification_type=notification_type,
service_id=service_id
)
current_app.logger.info( current_app.logger.info(
f'create-nightly-notification-status-for-day task for {process_day} type {notification_type}: ' f'create-nightly-notification-status-for-day task for {process_day} type {notification_type}: '

View File

@@ -1,6 +1,5 @@
from datetime import datetime, time, timedelta from datetime import datetime, time, timedelta
from flask import current_app
from notifications_utils.timezones import convert_bst_to_utc from notifications_utils.timezones import convert_bst_to_utc
from sqlalchemy import Date, case, func from sqlalchemy import Date, case, func
from sqlalchemy.dialects.postgresql import insert from sqlalchemy.dialects.postgresql import insert
@@ -36,19 +35,15 @@ from app.utils import (
) )
def fetch_notification_status_for_day(process_day, notification_type): 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)) 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)) end_date = convert_bst_to_utc(datetime.combine(process_day + timedelta(days=1), time.min))
current_app.logger.info("Fetch ft_notification_status for {} to {}".format(start_date, end_date)) # query notifications or notification_history for the day, depending on their data retention
service = Service.query.get(service_id)
all_data_for_process_day = []
services = Service.query.all()
# for each service query notifications or notification_history for the day, depending on their data retention
for service in services:
table = get_notification_table_to_use(service, notification_type, process_day, has_delete_task_run=False) table = get_notification_table_to_use(service, notification_type, process_day, has_delete_task_run=False)
data_for_service_and_type = query_for_fact_status_data( return query_for_fact_status_data(
table=table, table=table,
start_date=start_date, start_date=start_date,
end_date=end_date, end_date=end_date,
@@ -56,10 +51,6 @@ def fetch_notification_status_for_day(process_day, notification_type):
service_id=service.id service_id=service.id
) )
all_data_for_process_day += data_for_service_and_type
return all_data_for_process_day
def query_for_fact_status_data(table, start_date, end_date, notification_type, service_id): def query_for_fact_status_data(table, start_date, end_date, notification_type, service_id):
query = db.session.query( query = db.session.query(
@@ -86,18 +77,19 @@ def query_for_fact_status_data(table, start_date, end_date, notification_type, s
@autocommit @autocommit
def update_fact_notification_status(data, process_day, notification_type): def update_fact_notification_status(transit_data, process_day, notification_type, service_id):
table = FactNotificationStatus.__table__ table = FactNotificationStatus.__table__
FactNotificationStatus.query.filter( FactNotificationStatus.query.filter(
FactNotificationStatus.bst_date == process_day, FactNotificationStatus.bst_date == process_day,
FactNotificationStatus.notification_type == notification_type FactNotificationStatus.notification_type == notification_type,
FactNotificationStatus.service_id == service_id,
).delete() ).delete()
for row in data: for row in transit_data:
stmt = insert(table).values( stmt = insert(table).values(
bst_date=process_day, bst_date=process_day,
template_id=row.template_id, template_id=row.template_id,
service_id=row.service_id, service_id=service_id,
job_id=row.job_id, job_id=row.job_id,
notification_type=notification_type, notification_type=notification_type,
key_type=row.key_type, key_type=row.key_type,