You can configure scheduled events to NONE or SPECIFIED on one of the servers. When specified you need to add a comma separated list of scheduled event names in the yaml file (I asumed you use runtimes for this).
Mind you that turning of scheduled events might have impact on the HA if those scheduled events need to be running all the time. If that is the case and the server running the scheduled event goes down, your system becomes unavailable (or partly unavailable).
May be this trick could work (i forgot that the servers are offcourse unaware of which microflows are triggered). Create a entity with just one boolean in it. If this boolean is set to true a lDAP sync is in progress. Now use the community commons acquire lock in the scheduled event action. Only one of the servers will get the lock and the other server will detect that this record is already locked. If lock is detected break the microflow. If lock is acquired run the microflow that is really doing the sync.
I have not tested this so really do test this, but I think this should work because the lock is on the database level en thus server independent.