Bug: Stripe webhook dispatches billing task with None invoice_id, causing silent Celery failures

[shivam-pawar-7217]

Description

In stripe_webhook_view, when a payment_intent.succeeded event is received, invoice_id is extracted from Stripe's payment_intent metadata. However, there is no guard before dispatching the Celery task if invoice_id is absent from the metadata (which is valid for non-billing PaymentIntents or Stripe dashboard test events), the task is dispatched with invoice_id=None.

Affected Code

# app/eventyay/api/views/stripe.py
if event.type == 'payment_intent.succeeded':
   invoice_id = event.data.object.get('metadata', {}).get('invoice_id')
    update_billing_invoice_information.delay(invoice_id=invoice_id)  # dispatched even if None

Impact

• Silent failure: The update_billing_invoice_information Celery task catches None and returns without raising an exception. Consequently, the task shows SUCCESS in monitoring even though nothing was updated, hiding billing processing errors.
• Wasted resources: A Celery worker is scheduled and consumed for every non-billing payment_intent.succeeded event, even those that can never result in a billing update.
• Partial idempotency gap: The update() call in the task does not filter by current status, meaning rapid Stripe retries can produce multiple overlapping writes, resulting in an incorrect paid_datetime value for audit purposes.

Proposed Fix

• Add a None guard in the view before dispatching the task (if invoice_id:).
• Add a status=STATUS_PENDING filter in the Celery task's update() call to ensure exactly-once semantics and prevent concurrent workers from overwriting paid_datetime.