In the below dataset, there are two different ISPs for the user from their usual ones.
NordVPN for John and Quadranet for Jill – but this search using anomalidetection is detecting only john’s but not Jill’s.
Any idea why and what is the better way to detect the ISP outlier?
source="isp_data2.csv" index="test" sourcetype="csv"
| anomalydetection "ISP" "EmailAddress" action=annotate
| eval isOutlier = if(probable_cause != "", "1", "0")
| table "ISP" "EmailAddress", probable_cause, isOutlier
| sort 100000 probable_cause
Hi @ashishmgupta,
As a quick fix, try:
| anomalydetection "ISP" "EmailAddress" action=annotate cutoff=false
You seem to be missing the filter command
| where isOutlier=1