This incident has been resolved.
We implemented a fix and are currently monitoring the result.
Existing buckets are unaffected. All object operations (uploads, downloads, reads, writes, deletes) continue to work normally.