i was wondering in general what would be a good batch size to use for flushToIndexes.

case A.) you have 2 million items.

case B.) you have 1 billion items.

case C.) you have 5 billion items.

thanks

Hi,
it really depends on the amount of RAM available to your JVM and the complexity of your entities, it is really different if you index just one simple field or if you have several collections to index as embedded fields.
You should experiment, and probably use a profiler to help identify the bottleneck.

Generally you want to have the batch size as high as possible, but without running out of memory.