We are happy to announce that we have migrated our individual spam services on each server to a clustered setup allowing our spam system to learn about valid and invalid emails being sent to our system.
In the past we setup anti spam mechanisms on each individual server (SpamAssassin), learning what is and what isnt spam, to help protect our users.
We also implemented various security measures like RBL's (Real time block lists) URIBL (URL Block Lists) and many other techniques to help stop spam.
While this worked well in the past; it would sometimes deny valid spam from networks that were unreliable like GoDaddy, HostMonster and many other low end hosting providers.
These providers did not offer proper guidance or setup to their customers and would cause massive amounts of spam to be generated from their network. This would cause valid and non-spammy clients to suffer the consequences.
In an effort to not alienate these providers we implemented a "accept all" technique and scan individual emails. This would allow us to compare all emails sent and received and give them a score.
Do to the large amount of emails we send / receive we decided to learn on a global scale and unify our spam system.
We have launched RSPAMD internally for all systems, this service will process spam faster and more efficiently then SpamAssassin.
As this is a new system we still rely on SpamAssassin as a backup incase of issues and during our global testing.
If you have any issues please open a ticket with the time/date and email address you used to send or received email that could of been rejected.
Monday, December 5, 2016