I've been working on elasticsearch which always gets 100% cpu utilization while memory and heap utilization are pretty low (around 20-25% only). The main search being triggered (approximately 80% of the time) is a summation of a field filtered by date ranges; and batch indexing (approximately 20% of the time with only 200 records per transaction).
Could someone help me identify the cause of the high cpu utilization?
Environment Details per node:
Elasticsearch Version: 2.3.1
Java Version: 1.8.0_74
My cluster details (all default) are as follows:
"number_of_nodes" : 3,
"number_of_data_nodes" : 3,
"active_primary_shards" : 5,
"active_shards_percent_as_number" : 100.0
Here's a link to the hot-threads
Here's a link to the query