If a ReplicatorInstance leader gets too many logData() requests in a short period of time, then for whatever reason its followers will time out. Then the followers will stage an election, and all the requests the leader is working on will be useless. On my machine, with a quorum of three replicators backed by OLogs, this happens around 200k "simultaneous" logData() requests. More investigation is needed to understand why this is happening.
If a ReplicatorInstance leader gets too many logData() requests in a short period of time, then for whatever reason its followers will time out. Then the followers will stage an election, and all the requests the leader is working on will be useless. On my machine, with a quorum of three replicators backed by OLogs, this happens around 200k "simultaneous" logData() requests. More investigation is needed to understand why this is happening.