Sorry folks, this is a tough one to put into words, even worse because English is my second language, but bear with me please.
I have the following situation in hand and am not sure if it will work as expected:
The object I have on ElasticSearch has nested objects with different input streams, so that means that several of my documents receive lots of updates per minute.
These updates are mostly done on the nested docs.
To minimize the impact that this amount of operations does on my cluster's CPU I setup the refresh interval to 30 seconds and also made most of the operations to be done in bulk.
However I am worried that this approach may lead to a troublesome outcome: If I send two operations inside the same bulk changing different fields of a nested document, will the next GET receive the fully updated nested document?