Answer
Check the status of the specific Application Message responsible for triggering the sync — if it's ended up in a Finished state instead of cycling continuously through Waiting, the sync will stop actually triggering even though the scheduled job itself keeps running normally. Querying fndcn_application_message_tab for that message and updating its state back to 'Waiting' (and clearing any stale ERROR_TEXT) restores its normal continuous-trigger behavior. Check first whether there's a supported client action to reset a stuck Application Message before resorting to a direct table update, and take a backup/export of the row first regardless.