At 03:00 UTC on November 1, 2023, we observed degraded performance with the Object Storage Service in our Newark Datacenter after experiencing a very large amount of inbound traffic. This resulted in customers experiencing 503 errors and connection timeouts. By 09:33 UTC, we identified and restricted the problematic traffic and restarted key infrastructure components. These actions resolved the immediate impact of the issue.
To provide further resiliency against future impacts, we are upgrading the Object Storage infrastructure to higher-performance hardware and have made improvements to our monitoring systems.