Ben Thorner a7b39a930c Add comment to explain status aggregation approach ...

This relates to the performance optimisation work we're doing [1].
Before optimising the task, it's worth asking if we can do less -
the comment explains why it has to be this way.

Some references to back up the comment:

- We do status updates in either table [2].
- We don't allow duplicate receipts for emails [3].
- We don't allow duplicate receipts for SMS [4].
- We don't expect duplicate receipts for letters.

This is something we would need to revisit if we want to support
additional status updates - we could reject based on the age of the
notification, rather than the status.

[1]: https://github.com/alphagov/notifications-api/pull/3417
[2]: 20ead82463/app/dao/notifications_dao.py (L538)
[3]: 20ead82463/app/celery/process_ses_receipts_tasks.py (L58)
[4]: 20ead82463/app/dao/notifications_dao.py (L129-L135)

2022-01-10 18:15:54 +00:00

..

__init__.py

…

broadcast_message_tasks.py

…

letters_pdf_tasks.py

Log activity on all periodic Celery tasks

2021-11-17 09:48:03 +00:00

nightly_tasks.py

update log messages in delete task.

2021-12-14 15:24:35 +00:00

process_ses_receipts_tasks.py

DRY-up conditionally creating callback tasks

2021-12-06 14:11:34 +00:00

process_sms_client_response_tasks.py

DRY-up conditionally creating callback tasks

2021-12-06 14:11:34 +00:00

provider_tasks.py

…

reporting_tasks.py

Add comment to explain status aggregation approach

2022-01-10 18:15:54 +00:00

research_mode_tasks.py

…

scheduled_tasks.py

Remove alert for email / sms in created

2021-12-06 14:11:42 +00:00

service_callback_tasks.py

…

tasks.py

…