This stopped months ago and we didn't notice (we got a failmail at the time but I guess we didn't do anything).
Can we add monitoring either from the logs or from the civicrm_queue_item table?
The logs look like [2026-05-03T17:46:28.203856+00:00] civicrm.batch_merge.INFO: deduping 100 contacts from date 2026-05-01 13:58:07 to date 2026-05-01 14:58:07 {"limit":100,"start_date":"2026-05-01 13:58:07","to_date":"2026-05-01 14:58:07"} [] in CiviCRM.batch_merge.ABC123.log.N
Table row: | 17642306 | batch_merge | 0 | 2026-08-08 00:15:49 | 2026-08-08 01:15:49 | a:6:{s:15:"start_timestamp";s:19:"2026-05-01 14:00:00";s:9:"increment";s:10:"60 minutes";s:11:"batch_limit";i:100;s:8:"group_id";N;s:13:"rule_group_id";N;s:18:"minimum_contact_id";i:20561314;} | 1 |
In normal conditions, we should get one run per hour.

We may want to think about this in the context of broader monitoring for coworker tasks.