I’m troubleshooting some high usages in CPU by fleet/osquery in some of our hosts, and I’ve noticed that the config refresh takes quite some CPU (10~20%) every config interval.
With that in mind, I was wondering, since I already use orbit to distribute configs as set in Agent Options, do I need this to use OSQuery’s config plugin? Is it used by any means by orbit? If I set
to 0, will I still be able to set those configs via UI, or orbit only sets command_line configs, as opposed to query ones?
03/10/2023, 7:08 PM
Hi @Jesus Santos! The majority of options set in agent options are actually picked up by osquery using the config plugin. Only the command line flags are managed by Orbit. Can you take a look at the Orbit logs on one of the affected hosts to see if there's any information there that might help diagnose the increased usage?
03/13/2023, 12:17 PM
I increased the pull interval of configs by OSQuery, I think it will help in the overall consumption in this case.
If anyone else complain of this I’ll definitely take a look there (was focusing on the OSQuery logs itself)
Thanks for the clarification!