The sample section controls the parameters of the virtual spamtrap feature. The virtual spamtrap network collects messages that would normally be discarded based on the reputation of the message source (IP) when those messages do not already match pattern rules. In other words, the virtual spamtrap collects random samples of new spam from known spam sources. This feature is easily disabled if there are security concerns.

<sample on-off='on' probability='0.8' grab-one-in='5' passthrough='no' passthrough-symbol='0'/>

The on-off='on' attribute enables and on-off='off' disables this feature.

The probability='0.8' attribute establishes the minimum sample probability figure as 0.8. By default this coveres the entire black range.

The grab-one-in='5' attribute establishes the sample rate as one out of 5 candidate messages. By default every 5th message that fails to match a pattern rule and has a source IP that falls in the black range will be collected for the virtual spamtrap.

Passthrough Feature (special spamtrap processing)

The passthrough='no' disables and passthrough='yes' enables the sample pass-through feature. When enabled the feature prevents the message from being delivered to the virtual spamtrap network and instead causes the message scan to produce a specific result code. This feature is generally used only for spamtrap processing where the sampled messages should be delivered through to the spamtrap system. We use this feature internally for precisely that purpose.

The passthrough-symbol='0' establishes the symbolic result code for a pass-through sample as 0 by default. The default allows the message to look "clean" so that no action is taken to adjust its delivery. In special cases, a unique result code could be used so that when that code is returned the message can be delivered in some special way.

Please email support@armresearch.com with any questions.