Ocean Security
Resolved
The issue has been fully contained, and we have been monitoring the system for over 24 hours with no further issues observed.
Monitoring
Starting at approximately 16:00 IDT (13:00 UTC) on September 28, some inbound emails experienced delayed analysis due to failures with an upstream AI model. During this time, some emails were delivered without a complete threat verdict, and some user-reported phishing emails were not analyzed and did not receive a response.
We deployed a fix, and our analysis pipeline returned to normal operation. We re-analyzed the affected emails, and any emails identified as spam or malicious were remediated according to each organization’s policies. User-reported phishing emails that had failed analysis were also reprocessed.
We continued to monitor the system and provided another update once reprocessing was complete.