They seem to be blaming a 6h+ outage on a bug in Postgres but they don't reference which bug specifically. Anyone have any idea what they might be referring to?