By default, the system generates a watermark label at the bottom of each notification email to allow matching incoming email to existing records. Each watermark includes a random 20-character string that makes it unique.

Starting with the Jakarta release, the system automatically generates randomized watermarks for notification emails in base systems. The random 20-character string reduces the possibility of a watermark being guessed or coincidentally matching the watermark of an email from another instance.

Note: If you are upgrading from a release before Jakarta, random watermark support is optional and requires the Random Watermark Support plugin to be activated.

Watermark format

The email watermark always begins with "Ref:" to identify the label as a watermark. After this identifier, the default label is 31 characters in length and consists of:
  • Customizable prefix — The default prefix is MSG.
  • Auto-numbered identifier — The numeric string identifying the source record, such as incident, problem, or change request.
  • An underscore character followed by a random 20-character string
Figure 1. Randomized watermark example
Randomized watermark format

When inbound emails are processed, the system matches random watermarks to the appropriate source records.

Watermark configuration

Watermarks are always generated, but you can configure them to:
  • Create a custom watermark prefix for each instance to prevent accidentally triggering events in the wrong instance.
  • Have custom prefix characters after MSG
  • Be hidden globally
  • Be omitted from individual email messages
If watermarks are omitted from email notifications, inbound email actions might not work properly. Without a watermark, the system processes inbound email messages as described in Criteria for matching email to inbound actions.
Note: Email clients that use the plain text version of the email still show the watermark.