Hi all,
Why the count of "Event per day" in the "Indexing audit" dashboard is not match with |tstats result?
Eg.
The number from "Event per day" in the "Indexing audit" dashboard:
index count
main 10000
The number from |tstats count where index=main by index:
index count
main 500
May I know which one is correct?
It's possible both are correct. Or neither is correct. The two use different methods to arrive at numbers that may or may not represent the same thing. Also, we don't know the time window of each search.
The Indexing Audit dashboard uses a lookup table which is populated by a saved search. That search pulls data from metrics.log, which is a sampling of events rather than a true count.
The tstats query is counting actual events present in the index at the time the query runs.
Assume the time window is same. But from my environment,
the count from indexing audit dashboard is always 4-5 times more than the result from |tstats. If I want to know the Event per day. Which method is more accurate?
tstats is a count of all events whereas the Indexing Audit dashboard counts a sample of events. I would say tstats is more accurate.
Thanks for your response. Can you explain more on the "Indexing Audit dashboard" like how it counts the event count exactly. As it count on sample of event then the count should lower than |tstats count? Then why my "Indexing Audit dashboard" always 4-5 times higher than |tstats count ?
The dashboard gets its data from a KVStore lookup (licensing_epd). If the data in that lookup is inaccurate then so will the dashboard. See the Audit - Events Per Day - Lookup Gen saved search for how the lookup is populated.