The increased errors during request creation were caused by an RDS Postgres failover, observed between 09:47–09:54. Service has since recovered.
