You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The prebuilt detection rules sometimes produce false positive alerts either during normal system operation or due to some popular software triggering them. Additionally, the highlighted fields sometimes do not include relevant information and could be improved.
When rule improvements have been reported in the support system, there have been very different responses, including:
A suggestion to add manual rule exceptions.
A suggestion to wait for Elastic 9.0 and modify the upstream rules.
A response that the suggestion has been passed to a "Product Management team" (which might or might not be different from the rule maintainers).
A response that the rule maintainers have evaluated the suggestion and rejected it (a fairly rare occurrence) or included it (in most of the cases so far).
As it seems highly desirable to have such improvements upstream-ed both from the customer (no need to maintain them, especially in multiple installations) and Elastic (all customers would benefit from the rules being moreuseful) point of view, seeking an official position on the best process here.
This is a companion issue to a case in the support system.
The text was updated successfully, but these errors were encountered:
@richlv, I'm currently planning to work with the support engineers in our next cycle to improve our knowledge bases to reduce such discrepancies. This feedback suggests that we’re heading in the right direction, I appreciate it.
From an effectiveness point of view, reporting FPs directly here in the repo or communicating it to us via community slack is probably the way that will take less time to get them to our attention, as we—the detection developers—are monitoring the repository and the Slack channel. In contrast, we only get to know about support tickets when escalated. The disadvantage is that there is no SLA in the repository, and we may miss issues for some time eventually.
The prebuilt detection rules sometimes produce false positive alerts either during normal system operation or due to some popular software triggering them. Additionally, the highlighted fields sometimes do not include relevant information and could be improved.
When rule improvements have been reported in the support system, there have been very different responses, including:
As it seems highly desirable to have such improvements upstream-ed both from the customer (no need to maintain them, especially in multiple installations) and Elastic (all customers would benefit from the rules being moreuseful) point of view, seeking an official position on the best process here.
This is a companion issue to a case in the support system.
The text was updated successfully, but these errors were encountered: