In some cases, it could be that the identity link import imports and saves the same user task more than once for a process instance. If this happens, future imports will fail because of how we handle the collection of user tasks during upserts, and future imports will be blocked. We actively handle the duplicate storing of tasks, but it is possible that some race condition or other scenario causes us to still store both. When this happens, we should handle the case on future upserts so that the data gets cleared and subsequent imports work as expected.
- When importing user tasks, we properly handle the case where an existing user task has been stored more than once. We remove the duplicate and import continues as expected