Please help me to create a search, where I need to detect any anomaly of any host sending excessive logs with compare to the last hour.
For eg, if a host is sending x events this hour and next hour if it will send x+25% then we should get a trigger.
Hi @rahul_mckc_splunk1
as @to4kawa hints, try something like this:
index=your_index earliest=-2@h latest=@h
| timechart count span=1h
| delta count AS diff
| reverse
| head 1
| eval perc=diff/count*100
| where perc>25
and run this search every hour.
Ciao.
Giuseppe
bin _time
, stats by _time
and eval
is useful to create threshold
finally, use where
to fire alert.