You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is there already an utility inside elasticsearch-java that helps accommodate these requirements?
As far as I understand the new architecture correctly the serialization happens at the very end of the chain. But there might be a place where an estimation could happen? I would not like to serialize my objects twice. Once while building the BulkOperation-List and once while sending the request to ElasticSearch.
The text was updated successfully, but these errors were encountered:
This is resolved by PR #474 that introduces a BinaryData type that can be used for bulk operation documents, and of which we can get the size in bytes.
This BinaryData can be either created directly, but it's also used transparently by the BulkIngester to evaluate the size of the bulk request.
Description
Our application creates a
Stream
of objects that we write with anIndexRequest
into ElasticSearch. The Stream is consumed in chunks.The chunk size is determined by a few metrics that helped us stabilized the index Prozess. These are:
bulkRequest.estimatedSizeInBytes()
)RamUsageEstimator.sizeOfMap(script.getParams())
)Is there already an utility inside
elasticsearch-java
that helps accommodate these requirements?As far as I understand the new architecture correctly the serialization happens at the very end of the chain. But there might be a place where an estimation could happen? I would not like to serialize my objects twice. Once while building the
BulkOperation
-List and once while sending the request to ElasticSearch.The text was updated successfully, but these errors were encountered: