we are finalizing our systems now and we're going to start the test license in August. Today after doing some more configurations inside Kibana, I've discovered the Host overview and Metrics overview which didn't get filled with information. So I've decided to turn on the metrics collection via the classic System Integration and it worked. Now after inspecting it for quite a long time, I've seen that the CPU usage on all hosts go up to like 2000% for 1 minute and then drop down to the percentage that they actually have. Did someone else had that issue too? If yes, what was your fix?
It's pretty difficult to diagnose any problem (or otherwise) based on the limited description.
Great but ... we dont know much at all about "your system". It's a 101 node cluster across 3 sites? You will be ingesting 50k documents per second? Its running on a cluster of Raspberry Pis ? You are using elasticsearch 6.8, 7.12, 8.18, 9.x ... ?
It doesn't really matter to know in great detail, but at least a brief description would be nice.
2000% is 20x ? You are seeing that spike where specifically ?
"all hosts" means "those forming your elasticsearch cluster"?
How often does it spike, once per hour/day/week, seemingly randomly, ... ?
What, if any, issue does it cause for the applications/users/... ?
Wildest of wild guesses, what sort of storage are you using, how is that storage connected to your cluster's nodes, ... ?
Hi, no help needed anymore, it miraculously healed itself
Will watch out for the next time to provide more information, thank you nonetheless @RainTown !
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.
