feat: failed notifications for assignments leave state as allocated

[iloveagent57]

Description

When the assignment email notification task fails, the corresponding assignment is now NOT moved to an errored state, but instead left in an allocated state. Furthermore, the admin-serializer for the assignment now indicates a "failed" learner state, and an error reason related to the failed notification attempt.
ENT-9037

Manual testing

Here's the new response from the serializer after an assignment was allocated, but the resulting notification task failed:

{
    "next": null,
    "previous": null,
    "count": 1,
    "num_pages": 1,
    "current_page": 1,
    "start": 0,
    "results": [
        {
            "uuid": "fdaeb538-2d9d-4875-87ca-2973754ed949",
            "assignment_configuration": "a9f97992-4894-4b80-9dde-b08ce3d7bff3",
            "learner_email": "test-failed-notification@example.com",
            "lms_user_id": null,
            "content_key": "edx+DemoX",
            "content_title": "Demonstration Course",
            "content_quantity": -14800,
            "state": "allocated",
            "transaction_uuid": null,
            "actions": [
                {
                    "created": "2024-09-03T19:42:04.877357Z",
                    "modified": "2024-09-03T19:42:04.877380Z",
                    "uuid": "fca9d15b-8a41-4353-a84e-5b9795f5f5a8",
                    "action_type": "notified",
                    "completed_at": null,
                    "error_reason": "email_error"
                },
                {
                    "created": "2024-09-03T19:42:04.898243Z",
                    "modified": "2024-09-03T19:42:04.898263Z",
                    "uuid": "525b77c0-9466-4faa-b729-055a8d73a40e",
                    "action_type": "learner_linked",
                    "completed_at": "2024-09-03T19:42:04.897991Z",
                    "error_reason": null
                }
            ],
            "earliest_possible_expiration": {
                "date": "2024-12-02T19:42:04Z",
                "reason": "NINETY_DAYS_PASSED"
            },
            "recent_action": {
                "action_type": "assigned",
                "timestamp": "2024-09-03T19:42:04Z"
            },
            "learner_state": "failed",
            "error_reason": {
                "action_type": "notified",
                "error_reason": "email_error"
            }
        }
    ],
    "learner_state_counts": [
        {
            "learner_state": "failed",
            "count": 1
        }
    ]
}

This leaves the admin-portal assignment table in a state where the assignment is visible to the admin with an indication of what has gone wrong and how to work around it:

Furthermore, since the assignment record's state is still allocated, the assigned course is visible and actionable to the learner from the learner portal MFE:

Merge checklist:

• ./manage.py makemigrations has been run
  • Note: This must be run if you modified any models.
    • It may or may not make a migration depending on exactly what you modified, but it should still be run.

Post merge:

• Ensure that your changes went out to the stage instance
• Deploy to prod instance

[pwnage101]

I think this should work (not creating a new learner_state specifically for this case) for keeping the assignment visible in the Assigned table, but I'm not 100% sure. The ticket description recommends a new learner state, but just wanted to make sure we confirmed the frontend doesn't actually need a new learner state.

[iloveagent57]

Yup, the frontend already respects the FAILED value, that's how we get the "Failed: Bad Email" in the admin-portal screenshot in the description.

[pwnage101]

nit: use constant instead of magic string.

[pwnage101]

the rest LGTM