Redshift is translating 'dbo' schema name in to 'public' schema name.

We have two Redshift clusters. One cluster updated to version 1.0.44126 over the weekend. After that, we started experiencing failures in some data extract processes. The error was: ERROR: schema "dbo" does not exist

Sure enough, after reviewing the queries we had inadvertently used dbo instead of public in the schema name. However, it worked for months prior to this latest Redshift version. I cannot find any information on version 1.0.44126. I cannot find any feature that might translate these schema names transparently.

Why is Redshift doing this?

I can reproduce the behavior on cluster version 1.0.43931.

create table public.test_table ( test integer );
select * from dbo.test_table;
drop table dbo.test_table;

The above all succeeds on cluster version 1.0.43931 and fails on 1.0.44126.