Avoid network overhead/bottleneck due to bulk returning all items


(André Cimander) #1

Is there a hidden request parameter for the bulk api to omit successful items in the /bulk response or get a sane response format (e.g. success count, error count, optional list of errord documents)?

When we send a 50MB bulk request to elasticsearch, we also receive a 50MB+ back in the response's item key, which is kind of crazy for an endpoint that is designed for high throughput.

We are resorting to ?filter_path=errors, but this breaks error identification and requires some ugly workarounds, and the coordinating node is still gathering all the items from the cluster.


(system) #2

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.