Splunk Enterprise Security

how do i calculate the average of logs received from a sourcetype over last 30 days and compare FOR EACH SOURCETYPE if percentage of dip is more than 70% in last 24 hours when compared to average logs for that particular sourcetype

Engager

| metadata type=sourcetypes index=* group by index

| search sourcetype=*
| where lastTime < (now() - 86400)

| eval Duration=tostring(now() - lastTime,"duration")

| search Duration="*"
| fields sourcetype lastTime Duration

| sort - lastTime

| eval lastTime = strftime(lastTime,"%Y/%m/%d %H:%M" )
| rex field=Duration "(?(\d+))+"

0 Karma

SplunkTrust
SplunkTrust

try something along those lines, search for the last 30 (or 31 days)

| tstats count as event_count where index=_internal by sourcetype _time span=24h
| eventstats avg(event_count) as days_avg by sourcetype
| eval delta = round(event_count - days_avg, 2)
| eval perc_change = round((event_count / days_avg) * 100, 2)
| where _time >= now() - 86400 
| eval alert = if(perc_change > 70 AND delta < 0,"ALERT","OK")

hope it helps

0 Karma

Engager

I like the concept.. the only thing which is varying is Event Count is less than days average.

Instead Event count should be number of logs received over a time (example- time picker lets say 30 days)
and Days_avg should be average of event count of 30 days divided by 30 (eventcount/30)

percentage change should be number of events received in last 24 hours should a dip of more than 70 percent when compared with Days_avg

0 Karma

Ultra Champion
 | tstats count as event_count where index=_internal by sourcetype _time span=24h
 | eval day=strftime(_time,"%d")
 | eventstats dc(eval(strftime(_time,"%d"))) as days
 | eventstats sum(event_count) as days_total by sourcetype
 | eval  days_avg= round(days_total / days, 2)
 | eval perc_change = round((event_count / days_avg) * 100, 2)
 | eval delta = round(event_count - days_avg, 2)
 | where _time >= now() - 86400 
 | eval alert = if(perc_change > 70 AND delta < 0,"ALERT","OK")

Thanks @adonio
I modify the query for @adonio .

0 Karma

Ultra Champion

what's the average of logs?

0 Karma

Engager

@to4kawa Overage count of logs(events) received over a specific time

0 Karma