Recent searches
No recent searches
Cloudmark - Spam Detection Option
Posted May 20, 2024
I would really suggest giving the option to not have Cloudmark/Spam detection; in my experience, most of the year this isn't a problem, but every couple of months I start accumulating massive numbers of falsely suspended tickets. I'm talking about big customers, names that everyone is familiar with, email domains that everyone is familiar with, and that isn't even flagged by our own email and security systems as spam or remote potential for spam.
This has been terrible for us as it drags delay, agents aren't as focused on suspended tickets and we pride ourselves on responding in minutes. And since we have our own spam filtering mechanisms/products, we are almost exclusively hurt/set back by the ZD one.
I've had multiple tickets with ZD support, and there's absolutely no reasoning for how it's working; I understood this is Cloudmark flipping their fingerprinting/hueristics every now and then and being extremely loose about what they flag as spam.
Not sure what worth this has and why it isn't a huge issue for a considerable amount of enterprise users here… Thanks.
2
3 comments
Sydney Neubauer
+1 agreed - we run into this all the time. We have to use the Shredder app to set up rules to autorecover as we have thousands of suspended tickets. It would be best if we can turn off the automatic spam filtering - we already have an external system that scans them anyways. If they pass that, then they are cleared to go into ZD.
1
Shawna James
Thank you for taking the time to provide us with your feedback. This has been logged for our PM team to review. We appreciate you following up here with the feedback after reaching out to support. For others who may be interested in this feature request, please add your support by upvoting this post and/or adding your use case to the comments below. Thank you again!
0
Support Team
Like OP, we are more than happy to opt out of the CloudMark system, but we cant.
This is not just a problem for suspended tickets. Individual replies will be suspended, depending on if the spam filter see a particular comment sourced from an IP that is on a SPAM list.
Example:
A ticket was opened yesterday and there is a few back and forth between the agent and customer gathering data on the issue. Then the ticket stops receiving updates from the customer. This has many legitimate reasons, sometimes the users find the problem and move on.
In this case the customer did reply. She used a legitimate email service (gmail.com), but the IP gmail.com used to send the email with the reply onto Zendesk was on a SPAM list (sometimes even the big guys get their IPs flagged).
Result: That reply never made it to the ticket. Rather it was suspended. We did not followup, since we had the last comment on the ticket waiting for them to respond.
We have found other times when there is multiple CCs on a ticket, but only replies from a specific email in CC are suspended. Again, its because their email service provider happened to use a dirty IP. We cannot control what IPs GMAIL chooses to use, nor can our customers.
Brought this up with Zendesk support, specifying that having the filter apply to new tickets could make sense, at least a false positive does not generate an auto reply so the customer knows something went wrong. But silently dropping individual comments on EXISTING tickets because the sending IP happens to be in an SBL is a fundamental flaw.
A spammer would need to match lots of attributes on an existing ticket to sneak into the conversation. Yet, the answer from support was:
Translation: until the internet cleans up all IP space, expect random comments to disappear. Make sure you review your suspension box every 15 minutes or before every reply.
Handing off the process of filtering your core product which is communication, to a third party and not having a way to let your customer base opt out / affect the process when the solution fails is bonkers. The customer base is vast and diverse. How did this implementation make it through review?
1