We have SSD's in use , unfortunately I am not collecting system metrics over time. So I do not have any CPU stats. But indexing is happening all the time so I think I can get you some stats from the system now .
you want CPU stats from all 3 nodes ?
previously things were dead slow , back then my shard had of 5 primary shards and 1 replica shard.
I have increase this to 9 primary and two replicas , I have got slight performance improvement.
40GB per hour is almost 1TB per day, which sounds like a lot for a 3 node cluster. I would recommend monitoring CPU usage and disk I/O, including iowait, to see if you can identify what is limiting performance when you are querying and indexing at the same time. It may be worthwhile looking at network traffic as well as this also potentially can be a bottleneck during heavy indexing.
What does your dashboards look like? How many visualizations? Any performance intensive aggregation types or filters being used?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.