By default, once you increase a second standby that participates in the replication quorum then pg_auto_failover increments its range-sync-standbys from zero to at least one to ensure that as a substitute, Once your Principal node is the only real one left, then all transactions which might be producing info are placed on-keep until a standby node