Tuning Data Indexing Performance

You can check the status of your Elasticsearch clusters by using Elasticsearch’s Marvel dashboard.

For instructions, see the documentation for Marvel on Elasticsearch’s website.

If you find that requests are placing too heavy a burden on Elasticsearch or if you want to increase or decrease the parallelism of requests, you can modify the values of two parameters in the /opt/mapr/conf/gateway.conf file in your source MapR cluster.

Tuning Parameters

gateway.es.cluster.maxClients
Specifies the number of connections between each MapR gateway and Elasticsearch clusters. The default number of connections per gateway is 1. If you want to increase the parallelism of requests from MapR-DB to Elasticsearch to try improving performance, experiment with increasing this value after checking that the Elasticsearch cluster involved has excess capacity to handle traffic from more clients.
gateway.es.request.maxsize.kb
Specifies the maximum size of requests that MapR gateways send to Elasticsearch to update indexes. The default value is 128 KB. The maximum value is 1 MB. If an Elasticsearch cluster experiences heavy traffic from other applications, reducing this value can reduce the load on Elasticsearch and improve performance.