According to the default routing rule of Elasticsearch, data in a bulk request is routed to different shards. When massive data is written and a large number of index shards exist, excessive internal requests forwarding may trigger bulk rejection. In a large-scale cluster, the long tail effect causes a high bulk request latency.

You can specify the index.bulk_routing configuration item to enable bulk route optimization. This function reduces the requests that need to be internally forwarded. For clusters containing a large number of shards, this function can improve write performance and reduce bulk rejection.

Procedure

Choose Clusters in the navigation pane. On the Clusters page, locate the target cluster, and click Access Kibana in the Operation column.
In the navigation tree on the left, choose Dev Tools.

On the Dev Tools page, run the following command:

PUT my_index 
{
  "settings": { 
    "index.bulk_routing": "local_pack"
  } 
}

**Table 1** Values of **index.bulk_routing**
Value	Description
default	The default routing mechanism of Elasticsearch is used. Records in a bulk request are split and routed independently.
pack	Data of a single bulk request is randomly routed to the same shard.
local_pack	The data of a single bulk request is routed to the local shard of the data node that receives the bulk request. If the node does not contain the corresponding index shard, the data is randomly routed to another node that contains the index shard.

Bulk Route Optimization

Procedure