Title
#general
s

seph

01/07/2019, 3:51 PM
It’s easy to run expensive queries. One thing to look at is the
osquery_schedule
table. https://blog.kolide.com/profiling-osquery-performance-with-kolide-cloud-8e01097469db has some info (and there are probably other blog posts about)
w

Woogs

01/07/2019, 11:40 PM
This post references wall_time, is wall_time the cumulative time the query has run across all executions? Or is this the time spent per each execution?
n

nickespo

01/08/2019, 3:01 PM
@Woogs hey, i wrote that blog post - wall time is based on the last run you can verify this by scheduling something expensive with a short interval, and querying
select name, executions, wall_time, user_time from osquery_schedule
you’ll see executions and user time increasing, and wall time should remain about the same
3:35 PM
woogs, after digging in the source, i was wrong on that, and i’ll have to update the blog - wall time is the difference between unixtimes, which can definitely be zero for fast queries
w

Woogs

01/08/2019, 9:32 PM
awesome! Thanks for taking a look, been trying to figure out how to interpret that data
n

nickespo

01/08/2019, 9:48 PM
i actually asked that same question a while ago and nobody picked up the nuance - i just wrote that knowing it was ‘milliseconds of execution time’
9:48 PM
thanks for calling it out 🙂