@Basavaraj wrote: Reference : https://zpettry.com/cybersecurity/splunk-queries-data-exfiltration/ ```Group timestamps by day. Changes the time for all events to 00:00:00``` | bucket _time span=1d ```Total each field with a name beginning with "bytes". Group the results by user, day, and source address.```
| stats sum(bytes*) as bytes* by user _time src_ip ```Find the most recent timestamp. Compute the average value and standard deviation of the bytes_out field.```
| eventstats max(_time) as maxtime avg(bytes_out) as avg_bytes_out stdev(bytes_out) as stdev_bytes_out ```Count the number of results. Compute the average number and standard deviation of bytes sent at least an hour before the most recent timestamp for each source address```. | eventstats count as num_data_samples avg(eval(if(_time < relative_time(maxtime, "@h"),bytes_out,null))) as per_source_avg_bytes_out stdev(eval(if(_time < relative_time(maxtime, "@h"),bytes_out,null))) as per_source_stdev_bytes_out by src_ip
```Keep only the results with at least 4 values for a source address. Also, the number of bytes sent must be more than 3 standard deviations above the average for that source and they had to be sent within the last hour.``` | where num_data_samples >=4 AND bytes_out > avg_bytes_out + 3 * stdev_bytes_out AND bytes_out > per_source_avg_bytes_out + 3 * per_source_stdev_bytes_out AND _time >= relative_time(maxtime, "@h")
```Compute how far the number of bytes sent is from average``` | eval num_standard_deviations_away_from_org_average = round(abs(bytes_out - avg_bytes_out) / stdev_bytes_out,2), num_standard_deviations_away_from_per_source_average = round(abs(bytes_out - per_source_avg_bytes_out) / per_source_stdev_bytes_out,2)
```Discard fields we don't need anymore``` | fields - maxtime per_source* avg* stdev*
... View more
