Skip to main content

Configuration

The general settings for the spam filter can be found in the spam-filter Sieve script. Each configuration parameter is defined using the Sieve scripting language's variable mechanism. To set a variable, the let keyword is employed, followed by the variable's name, and then its designated value.

For example:

let "ADD_HEADER_SPAM" "true";

By default, all SPAM filter settings are obtained from the spam-config in-memory lookup store.

Headers​

The results of the spam filter's analysis are conveyed to end users via two headers: X-Spam-Status and X-Spam-Result. These headers offer end users transparency and insights into the workings of the spam filter.

X-Spam-Status​

The ADD_HEADER_SPAM setting determines whether the X-Spam-Status header should be added to emails processed by the filter. The X-Spam-Status header indicates, in a straightforward manner, whether an email is classified as spam or not, typically displaying a 'Yes' or 'No', along with the email's final score.

For example:

let "ADD_HEADER_SPAM" "true";

By setting it to true, the header will be added to emails. If you wish to prevent this header from being added, you would set this to false.

X-Spam-Result​

The ADD_HEADER_SPAM_RESULT setting controls the addition of the X-Spam-Result header to emails. The X-Spam-Result header provides a detailed breakdown of the email's analysis, listing each tag applied by the spam filter and its associated score. This gives users a comprehensive view of why an email received its final spam score.

For example:

let "ADD_HEADER_SPAM_RESULT" "true";

When set to true, this header will be added, offering detailed insights into the email's classification. To omit this detailed header, you would change this setting to false.

Thresholds​

The threshold settings offer system administrators the ability to fine-tune how aggressively the spam filter acts upon incoming messages, providing flexibility in managing potential spam and ensuring that legitimate messages are delivered appropriately.

Spam Threshold​

The SCORE_SPAM_THRESHOLD setting defines the score threshold for marking messages as spam when the ADD_HEADER_SPAM feature is enabled. If an email's cumulative score exceeds this threshold, it will be classified as spam. The default setting is that any message with a score above 5.0 will be deemed as spam and the X-Spam-Status header will reflect this classification.

let "SCORE_SPAM_THRESHOLD" "5.0";

Discard Threshold​

The SCORE_DISCARD_THRESHOLD parameter determines the score threshold at which messages are discarded outright, meaning they will not be delivered to the recipient's mailbox. With the current setting at "0", it implies that this feature is essentially inactive, as it's unlikely for a message to have a score of 0 or more and be discarded based solely on the spam filter's analysis. However, if administrators wish to activate this feature, they can adjust the value to a more practical threshold.

let "SCORE_DISCARD_THRESHOLD" "6.0";

Reject Threshold​

The SCORE_REJECT_THRESHOLD parameter sets the score threshold for outright rejecting incoming messages. When a message's score surpasses this threshold, it will be rejected, and the sending server will typically receive a notification of this rejection. As with the discard threshold, the current setting of "0" indicates that this feature is not actively rejecting any messages based on their spam score. Administrators can modify this threshold if they want the filter to start rejecting messages at a specific score.

let "SCORE_REJECT_THRESHOLD" "12.0";

Directory​

The DOMAIN_DIRECTORY setting specifies which directory to use when looking up local domain names. The primary reason for this configuration is to optimize network resources. When the system identifies a domain as local, it can bypass certain checks like DNSBL (DNS-based Block Lists) or phishing verifications. These checks are typically more relevant for external or unfamiliar domains. By avoiding unnecessary checks on local domains, the system can process emails faster and more efficiently, reducing network overhead and enhancing performance.

Example:

let "DOMAIN_DIRECTORY" "'default'";

If left unspecified, the default directory will be used.

Spam database​

The SPAM_DB setting specifies which lookup store to use as the spam filter database. This database is used to store sender reputation information, bayesian classifier models, greylist data, message reply tracking and other similar data.

Example:

let "SPAM_DB" "'redis'";

If left unspecified, the default lookup store will be used.

Deliver to Junk Mail folder​

Stalwart Mail Server server can be configured to automatically move spam messages to a designated Junk Mail folder. This is done by setting the spam.header.is-spam configuration attribute to the name of the header that the spam filter uses to identify spam messages. The header name is case-insensitive. If the header is present in an incoming message, the message is moved to the Junk Mail folder. If the header is not present, the message is delivered to the inbox.

For example:

[spam.header]
is-spam = "X-Spam-Status: Yes"