Seven years ago, I set up a web app that uses SES to send notification emails. I verified a sending domain. The app has run continuously since that time, sending well under 5000 emails/month.
About a year ago, we suddenly lost domain verification. It appeared that somehow the relevant DNS records weren't on the nameserver. We believed that the nameserver company must've somehow lost or deleted the records -- which made sense to us because they seemed dysfunctional generally. So we moved the DNS to be hosted at the name registrar (at GoDaddy).
I recently got an email from AWS about new verification requirements. Our throughput isn't high enough for it to apply to us, but the email happened to casually mention that we didn't have any verified domains. A little alarmed, I go into SES and find that our verification was again lost, and we'd been sandboxed!
I've fixed it now, but I'm trying to understand why this keeps happening. When I went in to fix, the DNS had an existing TXT record with an "amazonses" value. Yet the records SES instructed me to add today were three CNAME's. This suggests AWS has been changing the specific record types required for verification -- is this the case? Was I notified of a changing requirement and somehow missed it?
AWS 官方已更新 7 個月前
