Multiple folks over on the Mailop list are reporting that Microsoft OLC (Outlook.com/Hotmail/etc)'s IPv6 inbound mail servers are deferring inbound mail delivery attempts with "451 4.7.500 Server busy. Please try again later" errors. The fix seems to be to stop trying to send it over IPv6 and send the mail to any IPv4 MX record instead, and then your queues will drain successfully.
What domains are affected? Likely all domains I've listed here as handled by Microsoft, and only if you send over IPv6, and only if the recipient domain is a Microsoft-hosted domain that has an MX record with a hostname that maps to an IPv6 address.
If you're sending mail using an ESP or CRM platform, you're probably not affected by this. Big mail sending platforms, especially US-based ones, almost exclusively use IPv4 IP addresses, not IPv6.
Why is this affecting mail over IPv6 only? Nothing has been confirmed, but I would theorize that Microsoft likely has separate pods or groups of servers, including one that services IPv6, and perhaps only that group of servers hosting the IPv6 service is currently unhealthy.
I'm not affected, because I personally have disabled my IPv6 interface long ago, to prevent issues sending to Gmail. I'm sure email over IPv6 works fine for some folks, but it's too bleeding edge for my tastes; I'll let others blaze that new frontier.
Finally at around 8:09pm US EDT on Oct 27, 2022, Microsoft removed the AAAA IPv6 records from the hostname hotmail-com.olc.protection.outlook.com and I can confirm that the switch back to IPv4 got all of our emails finally delivered. Hopefully Microsoft can do a better job of rolling out IPv6 for email!
ReplyDelete