Hi
Suppose I have this log source here:
index=main
sourcetype=pan
host=pa3250
It generates a massive amount of logs daily. I know sometime within the last 20 days it stopped sending traffic. What's the best search query to help me identify the day that logs stopped coming in?
You could run that more efficiently using tstats:
| tstats count where index=main sourcetype=pan host=pa3250 by _time
| bin _time span=1d
especially when running over longer periods of time.
FYI.
Yes there it is. Was trying to get the query working with TSTATS. That's much better.
Does the count indicate the # of events that came in?
Also, what do the values in the "count" indicate? Is it how many files came in?
Hi impurush,
Thanks. I'm getting close with that one. It outputs to a table with _time and count. How would I also throw in the "host" in that table so that I can run it for a bunch of equipment in the same index and sourcetype
Just remove the host=xyz from the tstats and add it to the by clause, i.e. by _time, host.
And yes, the count is the number of events received by host. If you want the number of unique file sources, do a count by source (which is the filename, if it was a monitored file).
Hi @verifi81,
index=main sourcetype=pan host=pa3250|timechart span=1d count
You can run the above query for the last 30 days and see the visualization as a line graph, then see when it stopped.
And if you want to exact time, select the data when it stopped and change the span=1h to see when it exactly stopped.
Hi @s2_splunk,
Thank you for the valuable suggestion, Indeed, the tstats is very fast than the normal query which I provided.
Hi @verifi81, Please use the below query to include the host too.
| tstats count where index=main sourcetype=pan host=pa3250 by _time,host span=1d
You could run that more efficiently using tstats:
| tstats count where index=main sourcetype=pan host=pa3250 by _time
| bin _time span=1d
especially when running over longer periods of time.
FYI.