Performance tuning for indexing between using dedicated coordinator code and using data nodes directly

In my use case, I have a heavy load to index data continuously. I configured my data nodes with quite big CPU/Memory and using SSD. Also I have 3 dedicated coordinator nodes. At this point, both data read and write are through the coordinator nodes. I am wondering whether I should use data nodes directly when indexing node, but just use coordinator nodes for kibana to read data from. Could someone explain what the tradeoffs here I need to consider about?

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