Just wanted to update you all on this issue, what we have done to mitigate it, and what we are doing in future to reduce the disruption.
A brief history...
As a brand new IP address and mail server, we did have a few teething problems not long after we launched XenForo Cloud. To potentially correct
@Paul B, at least from my memory, I don't think we got blocked by Microsoft more than once or twice before. And for the most part, these were simply reputation things, because we were a new server, and it has been mostly working fine for a little while.
What I think has happened, because it affected the mail server we use here
and the mail server we use for Cloud customers, is Linode, the hosting provider we use for those servers, may have had its entire AS ranges blocked.
We became aware of the XF.com issue just before Christmas. This was reported immediately, but Microsoft denied there was an issue. The email said we could reply to add further details, which I did. It wasn't until after Christmas that I became aware that the Cloud mail server was affected too. So I again replied to that previous email. Despite assurances in the email we received, these emails were seemingly ignored.
We submitted a fresh request and this was eventually actioned and resolved the issue for Cloud customers on January 8th. Annoyingly, Microsoft were still denying there was anything blocking XF.com emails, but this contradicts the exact bounce messages they were sending which explicitly states they are
In terms of what we're doing to reduce disruption in the future:
1. We're building two new mail servers to replace the existing ones
2. Each of these servers are going to have multiple public facing IP addresses
3. The new server will send mail using any one of the IP addresses configured in our "pool"
4. The new software we have set up is significantly better in helping us surface errors that may be affecting deliverability
This has been rolled out to XF.com already which means we're now successfully delivering messages to most email providers, though we will be monitoring closely for any bounces that may indicate blocks we need to address. We now have a pool of multiple IP addresses and we can effectively remove IP addresses that might be blocked from that pool and either restore them once the block is resolved or replace them with new IP addresses entirely. This is now as simple as assigning a new IP address, configuring it in our IP pool and updating our SPF records. We can also assign a priority to each IP address so new IP addresses can be configured to be less likely to be used while they "warm up". The fact that we have multiple IPs also means any significant volume of email is now less "alarming".
If the next day or two goes well, we can roll these changes out to Cloud customers too. Something which should be mostly transparent without any significant disruption to existing sites.
In summary: For now, the issues should be resolved, and in future they should be less likely to be as disruptive.
