When the connector decides to increment/decrement the batch size used for LogMining, the step is currently configured to use currentSize + minSize, which will create a very slow climb to max when the min is configured too small.

I propose we introduce a new log.mining.batch.size.increment setting, and this will control how fast/slow the batch calculation climbs. The current min/max will define the range while the default defines the starting value in the range.