These are chat archives for RBMHTechnology/eventuate
Global-scale event sourcing and event collaboration with causal consistency
replayBatchSize is only relevant when the writer catches up from a (long) history of events and is used for backpressure. Once the writer has caught up
replayBatchSize is not relevant.
Regarding batch sizes, the
replayBatchSize is an upper bound. If the writer currently doesn't have a write in progress and an event is consumed, write() is triggered immediately. See batching for details.
Regarding write() "called twice": do you have events that don't generate write batches (e.g. empty matches or whatever)?
Regarding processing, every writer (like every actor) does serial processing and multiple writers are running concurrently. The batch size settings are per writer instance.
Hope that helps.