We are experiencing intermittent -32500 (SQL failed) errors via the API when running parallel provisioning scripts and JIT AD syncing. Currently, this issue is impacting a client's environment with over 1,000 users.

While initially thought to be exclusively related to host.create, we observed this happening with user provisioning/updates as well ("cannot add/update user").

The issue seems to be a race condition highly similar to the logic discussed in ZBX-27448. When multiple parallel processes attempt to create entities with uniquely new host-group/user-group combinations, the processes attempt to "reserve" the same future IDs for the new group-set's hash. This results in a clash during insertBatch() within createHgSets(), ultimately failing when inserting related records into the permission table because the parent hgset record hasn't been successfully committed yet.

The API intermittently returns a -32500 (SQL failed) error.

Looking at the frontend debug logs (API response debug), the following foreign key constraint failure is exposed:

cannot add or update a child row: a foreign key constraint fails (`zabbix`.`permission`, CONSTRAINT `c_permission_2` FOREIGN KEY (`hgsetid`) REFERENCES `hgset` (`hgsetid`) ON DELETE CASCADE)