Description:

The issue involves Confluence not sending delete events correctly to the data lake, leading to inconsistencies in data representation. The investigation revealed the following sequence of events:

Space Creation:

Two events are sent to the data lake upon space creation.

Space Deletion (Sent to Trash):

A deletion event is sent to the data lake, marking the space as deleted.

Issue: Confluence should send a real-time record to mark the space as "trash" instead of "deleted."

Bootstrap Trigger:

A real-time event is sent to the data lake before the bootstrap event.

Observation: This behavior is consistent with customer instances.

Permanent Deletion from Trash:

No delete event is sent to the data lake.

Issue: Confluence should send a real-time record to mark the space as "deleted" (i.e., send a tombstone record).

Subsequent Bootstrap Trigger:

No event is sent for the space, triggering Plato garbage collection to delete the row.

Expected Behavior:

• Confluence should send a real-time record to mark spaces as "trash" when moved to trash.

• Upon permanent deletion, Confluence should send a real-time record to mark spaces as "deleted" (tombstone record).

• Ensure consistent event handling during bootstrap processes.

Steps to Reproduce:

• Create a space in Confluence.

• Delete the space (send to trash).

• Trigger a bootstrap.

• Permanently delete the space from trash.

• Trigger another bootstrap.

Impact:

This issue affects data integrity and consistency in the data lake, potentially leading to incorrect data representation and processing.