Were you performing any queries while this data was collected? Can you perform the same search a few times that took 82 seconds? Also, can you attach the search that you're performing?
@pickypg, I wasn't performing any search at that time. I just restarted the test cluster.
This is the search i was performing on the main cluster which took 82 seconds:
curl -XGET "http://localhost:9200/op-real-*/_search" -v -d'
{
"query": {
"match": {
"message": "/open-service-locator/api/v2/"
}
}}'
I performed the search few times and i am getting the results as the search data is being cached each time.
Elasticsearch does not cache queries unless "size": 0, but it would cache filters.
The only caching in play here would be the file system cache. If your performance was 82s without the file system cache for a regular query, then I have to wonder what is going on with the machine that is hosting the ES node.
@pickypg, my bad, i meant kibana was caching the results and providing them as i made same search multiple times from kibana.
If i perform the same search on ES:
first run took 89670 ms
second run took 140794 ms
third run took 107788 ms
fourth run took 140407 ms
fifth run took 98779 ms
I am hosting ES nodes on centos aws ec2 m4.large machines with 200 GB of EBS volumes attached. Currently the storage is 54% full.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.