Hi All,
we're tuning the Splunk App for Enterprise Security setup for one Customer and we're experiences a LOT of Notable Events for Correlational Search "Default account activity detected"generated also for not default user accounts, but for regular user accounts.
It seems that the "default_user_accounts" macro invoked by the rule doesn't look in the identity.csv file for default users, but in another "identities_expanded.csv" lookup file. The online docs for Splunk app for ES 2.4 mention a postprocess search that generate this csv but I could't find it.
Any idea on how to clean and set the default account detection correctly?
regards,
Marco
Luke,
thanks for this clarification, but how this "identities_expanded.csv" is build? Why an account not present in "identities.csv" the appears as "default account" in "identities_expanded.csv"?
Moreover, is there a way to "reset" it?
Thanks,
Marco
Hi Marco, this process is explained in the documentation link that Luke provided. Per my question, the category tag is the key here.
I expect that if you're seeing an unexpected category tag, then you've probably got multiple definitions of the account in question, one with and one without category="default".
Lastly, I would be remiss not to note that this system has been improved in the new ES 3.0; mind you, there is no fix for GIGO, so a misplaced category would still produce the unwanted result.
The "default_user_accounts" evaluates "identities.csv" but indirectly. To improve performance, the "identities.csv" is processed behind the scenes into "identities_expanded.csv".
Splunk will rebuild the identities_expanded.csv automatically. However, you can force it to run manually too. See the docs on "Changes to Assets and Identities"
also, the important thing is the identity's category... do the regular accounts have category=default set?