Does it therefore sound plausible that the Flink job with just 2 parallelism and therefore 2 writers would just keep sending requests to the same 2 nodes? Does the Java SDK for bulk indexing generally keep the same connections open for writes as long as possible?
100% it does.
Yes it does connection pooling and persistency.
1 Like
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.